output
bool 2
classes | input
stringlengths 345
2.91k
|
|---|---|
true |
Memory Optimization for Deep Networks
memory optimized training memory efficient training checkpointing deep network training
Deep learning is slowly, but steadily, hitting a memory bottleneck. While the tensor computation in top-of-the-line GPUs increased by $32\times$ over the last five years, the total available memory only grew by $2.5\times$. This prevents researchers from exploring larger architectures, as training large networks requires more memory for storing intermediate outputs. In this paper, we present MONeT, an automatic framework that minimizes both the memory footprint and computational overhead of deep networks. MONeT jointly optimizes the checkpointing schedule and the implementation of various operators. MONeT is able to outperform all prior hand-tuned operations as well as automated checkpointing. MONeT reduces the overall memory requirement by $3\times$ for various PyTorch models, with a 9-16$\%$ overhead in computation. For the same computation cost, MONeT requires 1.2-1.8$\times$ less memory than current state-of-the-art automated checkpointing frameworks. Our code will be made publicly available upon acceptance.
|
true |
Generative Scene Graph Networks
object-centric representations generative modeling scene generation variational autoencoders
Human perception excels at building compositional hierarchies of parts and objects from unlabeled scenes that help systematic generalization. Yet most work on generative scene modeling either ignores the part-whole relationship or assumes access to predefined part labels. In this paper, we propose Generative Scene Graph Networks (GSGNs), the first deep generative model that learns to discover the primitive parts and infer the part-whole relationship jointly from multi-object scenes without supervision and in an end-to-end trainable way. We formulate GSGN as a variational autoencoder in which the latent representation is a tree-structured probabilistic scene graph. The leaf nodes in the latent tree correspond to primitive parts, and the edges represent the symbolic pose variables required for recursively composing the parts into whole objects and then the full scene. This allows novel objects and scenes to be generated both by sampling from the prior and by manual configuration of the pose variables, as we do with graphics engines. We evaluate GSGN on datasets of scenes containing multiple compositional objects, including a challenging Compositional CLEVR dataset that we have developed. We show that GSGN is able to infer the latent scene graph, generalize out of the training regime, and improve data efficiency in downstream tasks.
|
true |
Teaching with Commentaries
learning to teach metalearning hypergradients
Effective training of deep neural networks can be challenging, and there remain many open questions on how to best learn these models. Recently developed methods to improve neural network training examine teaching: providing learned information during the training process to improve downstream model performance. In this paper, we take steps towards extending the scope of teaching. We propose a flexible teaching framework using commentaries, learned meta-information helpful for training on a particular task. We present gradient-based methods to learn commentaries, leveraging recent work on implicit differentiation for scalability. We explore diverse applications of commentaries, from weighting training examples, to parameterising label-dependent data augmentation policies, to representing attention masks that highlight salient image regions. We find that commentaries can improve training speed and/or performance, and provide insights about the dataset and training process. We also observe that commentaries generalise: they can be reused when training new models to obtain performance benefits, suggesting a use-case where commentaries are stored with a dataset and leveraged in future for improved model training.
|
true |
An Empirical Study on Prompt Compression for Large Language Models
prompt compression explanation faithfulness feature attribution robustness
Prompt engineering enables Large Language Models (LLMs) to perform a variety of tasks. However, lengthy prompts significantly increase computational complexity and economic costs. To address this issue, we study six prompt compression methods for LLMs, aiming to reduce prompt length while maintaining LLM response quality. In this paper, we present a comprehensive analysis covering aspects such as generation performance, model hallucinations, efficacy in multimodal tasks, word omission analysis, and more. We evaluate these methods across 13 datasets, including news, scientific articles, commonsense QA, math QA, long-context QA, and VQA datasets. Our experiments reveal that prompt compression has a greater impact on LLM performance in long contexts compared to short ones. In the Longbench evaluation, moderate compression even enhances LLM performance.
|
false |
On the Effectiveness of Deep Ensembles for Small Data Tasks
small data deep learning ensembles classification
Deep neural networks represent the gold standard for image classification.
However, they usually need large amounts of data to reach superior performance.
In this work, we focus on image classification problems with a few labeled examples per class and improve sample efficiency in the low data regime by using an ensemble of relatively small deep networks.
For the first time, our work broadly studies the existing concept of neural ensembling in small data domains, through an extensive validation using popular datasets and architectures.
We show that deep ensembling is a simple yet effective technique that outperforms current state-of-the-art approaches for learning from small datasets.
We compare different ensemble configurations to their deeper and wider competitors given a total fixed computational budget and provide empirical evidence of their advantage.
Furthermore, we investigate the effectiveness of different losses and show that their choice should be made considering different factors.
|
true |
Categorial Grammar Induction as a Compositionality Measure for Emergent Languages in Signaling Games
Emergent Communication Emergent Language Categorial Grammar Induction
This paper proposes a method to investigate the syntactic structure of emergent languages using categorial grammar induction.
Although the structural property of emergent languages is an important topic, little has been done on syntax and its relation to semantics.
Inspired by previous work on CCG induction for natural languages, we propose to induce categorial grammars from sentence-meaning pairs of emergent languages.
Since an emergent language born in a signaling game is represented as pairs of a message and meaning, it is straightforward to extract sentence-meaning pairs to feed to categorial grammar induction.
We also propose two compositionality measures that are based on induced grammars.
Our experimental results reveal that our measures can recognize compositionality.
While correlating with existing measure TopSim, our measures can gain more insights on the compositional structure of emergent languages from induced grammars.
|
true |
Towards Effective Discrimination Testing for Generative AI
algorithmic fairness generative AI language models text-to-image evaluation AI regulation algorithmic discrimination anti-discrimination law
Generative AI (GenAI) models present new challenges in regulating against discriminatory behavior. We argue that GenAI fairness research still has not met these challenges; instead, a significant gap remains between bias assessment methods and regulatory goals. This leads to ineffective regulation that can allow deployment of reportedly fair, yet actually discriminatory, GenAI systems. Towards remedying this problem, we connect the legal and technical literature around GenAI bias evaluation and identify areas of misalignment. Through four case studies, we demonstrate how this misalignment can result in discriminatory outcomes in real-world deployments, especially in adaptive or complex environments. We offer practical recommendations for improving discrimination testing to better align with regulatory goals and enhance the reliability of fairness assessments in the future.
|
true |
Representation Learning via Invariant Causal Mechanisms
Representation Learning Self-supervised Learning Contrastive Methods Causality
Self-supervised learning has emerged as a strategy to reduce the reliance on costly supervised signal by pretraining representations only using unlabeled data. These methods combine heuristic proxy classification tasks with data augmentations and have achieved significant success, but our theoretical understanding of this success remains limited. In this paper we analyze self-supervised representation learning using a causal framework. We show how data augmentations can be more effectively utilized through explicit invariance constraints on the proxy classifiers employed during pretraining. Based on this, we propose a novel self-supervised objective, Representation Learning via Invariant Causal Mechanisms (ReLIC), that enforces invariant prediction of proxy targets across augmentations through an invariance regularizer which yields improved generalization guarantees. Further, using causality we generalize contrastive learning, a particular kind of self-supervised method, and provide an alternative theoretical explanation for the success of these methods. Empirically, ReLIC significantly outperforms competing methods in terms of robustness and out-of-distribution generalization on ImageNet, while also significantly outperforming these methods on Atari achieving above human-level performance on 51 out of 57 games.
|
true |
Latent Convergent Cross Mapping
Causality Time Series Chaos Neural ODE Missing Values
Discovering causal structures of temporal processes is a major tool of scientific inquiry because it helps us better understand and explain the mechanisms driving a phenomenon of interest, thereby facilitating analysis, reasoning, and synthesis for such systems.
However, accurately inferring causal structures within a phenomenon based on observational data only is still an open problem. Indeed, this type of data usually consists in short time series with missing or noisy values for which causal inference is increasingly difficult. In this work, we propose a method to uncover causal relations in chaotic dynamical systems from short, noisy and sporadic time series (that is, incomplete observations at infrequent and irregular intervals) where the classical convergent cross mapping (CCM) fails. Our method works by learning a Neural ODE latent process modeling the state-space dynamics of the time series and by checking the existence of a continuous map between the resulting processes. We provide theoretical analysis and show empirically that Latent-CCM can reliably uncover the true causal pattern, unlike traditional methods.
|
true |
LLM Neurosurgeon: Targeted Knowledge Removal in LLMs using Sparse Autoencoders
LLM large language model sparse autoencoder autoencoder precise safety steering interpretable interpretability generative AI generative AI machine learning
Generative AI's widespread use has raised concerns about trust, safety, steerability, and interpretability. Existing solutions, like prompt engineering, fine-tuning, and reinforcement learning (e.g., RLHF, DPO), are often hard to iterate, computationally expensive, and rely heavily on dataset quality.
This paper introduces Neurosurgeon, an efficient procedure that uses sparse autoencoders to identify and remove specific topics from a language model’s internal representations. This approach offers precise control over model responses while maintaining overall behavior. Experiments on the Gemma 2-9B model show Neurosurgeon’s ability to reduce bias in targeted areas without altering the model’s core functionality.
|
false |
Which Model to Transfer? Finding the Needle in the Growing Haystack
model needle models haystack haystack transfer learning alternative scratch particular vision nlp
Transfer learning has been recently popularized as a data-efficient alternative to training models from scratch, in particular in vision and NLP where it provides a remarkably solid baseline. The emergence of rich model repositories, such as TensorFlow Hub, enables the practitioners and researchers to unleash the potential of these models across a wide range of downstream tasks. As these repositories keep growing exponentially, efficiently selecting a good model for the task at hand becomes paramount. We provide a formalization of this problem through a familiar notion of regret and introduce the predominant strategies, namely task-agnostic (e.g. picking the highest scoring ImageNet model) and task-aware search strategies (such as linear or kNN evaluation). We conduct a large-scale empirical study and show that both task-agnostic and task-aware methods can yield high regret. We then propose a simple and computationally efficient hybrid search strategy which outperforms the existing approaches. We highlight the practical benefits of the proposed solution on a set of 19 diverse vision tasks.
|
true |
Benchmarking Generative Latent Variable Models for Speech
generative models latent variable models variational autoencoder generative speech modelling benchmark likelihood phoneme recognition
Stochastic latent variable models (LVMs) achieve state-of-the-art performance on natural image generation but are still inferior to deterministic models on speech. In this paper, we develop a speech benchmark of popular temporal LVMs and compare them against state-of-the-art deterministic models. We report the likelihood, which is a much used metric in the image domain, but rarely, or incomparably, reported for speech models. To assess the quality of the learned representations, we also compare their usefulness for phoneme recognition. Finally, we adapt the Clockwork VAE, a state-of-the-art temporal LVM for video generation, to the speech domain. Despite being autoregressive only in latent space, we find that the Clockwork VAE can outperform previous LVMs and reduce the gap to deterministic models by using a hierarchy of latent variables.
|
true |
What makes a language easy to learn? A preregistered study on how systematic structure and community size affect language learnability
Language Learning Learning Bias Humans Compositionality Systematic Structure Generalization Convergence
Cross-linguistic differences in morphological complexity could have important consequences for language learning. Specifically, it is often assumed that languages with more regular, compositional, and transparent grammars are easier to learn by both children and adults. Moreover, it has been shown that such grammars are more likely to evolve in bigger communities. Together, this suggests that some languages are acquired faster than others, and that this advantage can be traced back to community size and to the degree of systematicity in the language. However, the causal relationship between systematic linguistic structure and language learnability has not been formally tested, despite its potential importance for theories on language evolution, second language learning, and the origin of linguistic diversity. In this pre-registered study, we experimentally tested the effects of community size and systematic structure on adult language learning. We compared the acquisition of different yet comparable artificial languages that were created by big or small groups in a previous communication experiment, which varied in their degree of systematic linguistic structure. We asked (a) whether more structured languages were easier to learn; and (b) whether languages created by the bigger groups were easier to learn. We found that highly systematic languages were learned faster and more accurately by adults, but that the relationship between language learnability and linguistic structure was typically non-linear: high systematicity was advantageous for learning, but learners did not benefit from partly or semi-structured languages. Community size did not affect learnability: languages that evolved in big and small groups were equally learnable, and there was no additional advantage for languages created by bigger groups beyond their degree of systematic structure. Furthermore, our results suggested that predictability is an important advantage of systematic structure: participants who learned more structured languages were better at generalizing these languages to new, unfamiliar meanings, and different participants who learned the same more structured languages were more likely to produce similar labels. That is, systematic structure may allow speakers to converge effortlessly, such that strangers can immediately understand each other.
|
false |
Behavior of Mini-Batch Optimization for Training Deep Neural Networks on Large Datasets
Stochastic gradient descent large datasets convex optimization parallelized computation deep neural networks
Stochastic Weight Averaging in Parallel (SWAP) is a method that enables the training of deep neural networks on large datasets using large mini-batch sizes while also not sacrificing good generalization behavior. The algorithm uses large mini-batches to calculate the approximate model weights. The final model weight is the average of refined weights through parallel small mini-batch training of the approximate weights. This post provides a summary of the paper presenting SWAP, in addition to providing simple explanations of both related and foundational concepts upon which SWAP builds. Important concepts such as convexity, generalizability and gradient descent are explained. Related approaches that aim to obtain good generalization properties for large mini batches, like, ensemble of model parameters and local updating methods are discussed. Model performance of SWAP is presented for the task of image classification by deep learning models using popular computer vision benchmark datasets, such as, CIFAR 10, CFAIR 100, and Image Net. Further possible improvements identified by the authors are elucidated upon and additional future directions are identified and explained by the post authors.
|
false |
Median DC for Sign Recovery: Privacy can be Achieved by Deterministic Algorithms
Median-of-means divide-and-conquer privacy sign recovery
Privacy-preserving data analysis becomes prevailing in recent years. It is a common sense in privacy literature that strict differential privacy can only be obtained by imposing additional randomness in the algorithm. In this paper, we study the problem of private sign recovery for sparse mean estimation and sparse linear regression in a distributed setup. By taking a coordinate-wise median among the reported local sign vectors, which can be referred to as a median divide-and-conquer (Med-DC) approach, we can recover the signs of the true parameter with a provable consistency guarantee. Moreover, without adding any extra randomness to the algorithm, our Med-DC method can protect data privacy with high probability. Simulation studies are conducted to demonstrate the effectiveness of our proposed method.
|
true |
Overfitting for Fun and Profit: Instance-Adaptive Data Compression
Neural data compression Learned compression Generative modeling Overfitting Finetuning Instance learning Instance adaptation Variational autoencoders Rate-distortion optimization Model compression Weight quantization
Neural data compression has been shown to outperform classical methods in terms of $RD$ performance, with results still improving rapidly.
At a high level, neural compression is based on an autoencoder that tries to reconstruct the input instance from a (quantized) latent representation, coupled with a prior that is used to losslessly compress these latents.
Due to limitations on model capacity and imperfect optimization and generalization, such models will suboptimally compress test data in general.
However, one of the great strengths of learned compression is that if the test-time data distribution is known and relatively low-entropy (e.g. a camera watching a static scene, a dash cam in an autonomous car, etc.), the model can easily be finetuned or adapted to this distribution, leading to improved $RD$ performance.
In this paper we take this concept to the extreme, adapting the full model to a single video, and sending model updates (quantized and compressed using a parameter-space prior) along with the latent representation. Unlike previous work, we finetune not only the encoder/latents but the entire model, and - during finetuning - take into account both the effect of model quantization and the additional costs incurred by sending the model updates. We evaluate an image compression model on I-frames (sampled at 2 fps) from videos of the Xiph dataset, and demonstrate that full-model adaptation improves $RD$ performance by ~1 dB, with respect to encoder-only finetuning.
|
true |
Torsional Diffusion for Molecular Conformer Generation
conformer generation diffusion models equivariance riemannian manifold
Diffusion-based generative models generate samples by mapping noise to data via the reversal of a diffusion process that typically consists of independent Gaussian noise in every data coordinate. This diffusion process is, however, not well suited to the fundamental task of molecular conformer generation where the degrees of freedom differentiating conformers lie mostly in torsion angles. We, therefore, propose Torsional Diffusion that generates conformers by leveraging the definition of a diffusion process over the space $\mathbb{T}^m$, a high dimensional torus representing torsion angles, and a $SE(3)$ equivariant model capable of accurately predicting the score over this process. Empirically, we demonstrate that our model outperforms state-of-the-art methods in terms of both diversity and accuracy of generated conformers, reducing the mean minimum RMSD by respectively 32% and 17%. When compared to Gaussian diffusion models, torsional diffusion enables significantly more accurate generation while performing two orders of magnitude fewer inference time-steps.
|
true |
SHREWD: Semantic Hierarchy Based Relational Embeddings For Weakly-Supervised Deep Hashing
Deep Hashing Content Base Image Retrieval Semantic image relations Weakly Supervised learning Representation Learning
Using class labels to represent class similarity is a typical approach to training deep hashing systems for retrieval; samples from the same or different classes take binary 1 or 0 similarity values. This similarity does not model the full rich knowledge of semantic relations that may be present between data points. In this work we build upon the idea of using semantic hierarchies to form distance metrics between all available sample labels; for example cat to dog has a smaller distance than cat to guitar. We combine this type of semantic distance into a loss function to promote similar distances between the deep neural network embeddings. We also introduce an empirical Kullback-Leibler divergence loss term to promote binarization and uniformity of the embeddings. We test the resulting SHREWD method and demonstrate improvements in hierarchical retrieval scores using compact, binary hash codes instead of real valued ones, and show that in a weakly supervised hashing setting we are able to learn competitively without explicitly relying on class labels, but instead on similarities between labels.
|
true |
On Representing Electronic Wave Functions with Sign Equivariant Neural Networks
Quantum Monte Carlo Schrödinger Equation Neural Network Wave Function Antisymmetric Functions Computational Physics Quantum Chemistry Neural Quantum States
Recent neural networks demonstrated impressively accurate approximations of electronic ground-state wave functions. Such neural networks typically consist of a permutation-equivariant neural network followed by a permutation-antisymmetric operation to enforce the electronic exchange symmetry. While accurate, such neural networks are computationally expensive. In this work, we explore the flipped approach, where we first compute antisymmetric quantities based on the electronic coordinates and then apply sign equivariant neural networks to preserve the antisymmetry. While this approach promises acceleration thanks to the lower-dimensional representation, we demonstrate that it reduces to a Jastrow factor, a commonly used permutation-invariant multiplicative factor in the wave function. Our empirical results support this further, finding little to no improvements over baselines. We conclude with neither theoretical nor empirical advantages of sign equivariant functions for representing electronic wave functions within the evaluation of this work.
|
false |
Meta-learning Transferable Representations with a Single Target Domain
transfer learning fine-tuning supervised transfer learning
Recent works found that fine-tuning and joint training---two popular approaches for transfer learning---do not always improve accuracy on downstream tasks. First, we aim to understand more about when and why fine-tuning and joint training can be suboptimal or even harmful for transfer learning. We design semi-synthetic datasets where the source task can be solved by either source-specific features or transferable features. We observe that (1) pre-training may not have incentive to learn transferable features and (2) joint training may simultaneously learn source-specific features and overfit to the target. Second, to improve over fine-tuning and joint training, we propose Meta Representation Learning MeRLin to learn transferable features. MeRLin meta-learns representations by ensuring that a head fit on top of the representations with target training data also performs well on target validation data. We also prove that MeRLin recovers the target ground-truth model with a quadratic neural net parameterization and a source distribution that contains both transferable and source-specific features. On the same distribution, pre-training and joint training provably fail to learn transferable features. MeRLin empirically outperforms previous state-of-the-art transfer learning algorithms on various real-world vision and NLP transfer learning benchmarks.
|
true |
Discovering Diverse Multi-Agent Strategic Behavior via Reward Randomization
strategic behavior multi-agent reinforcement learning reward randomization diverse strategies
We propose a simple, general and effective technique, Reward Randomization for discovering diverse strategic policies in complex multi-agent games. Combining reward randomization and policy gradient, we derive a new algorithm, Reward-Randomized Policy Gradient (RPG). RPG is able to discover a set of multiple distinctive human-interpretable strategies in challenging temporal trust dilemmas, including grid-world games and a real-world game Agar.io, where multiple equilibria exist but standard multi-agent policy gradient algorithms always converge to a fixed one with a sub-optimal payoff for every player even using state-of-the-art exploration techniques. Furthermore, with the set of diverse strategies from RPG, we can (1) achieve higher payoffs by fine-tuning the best policy from the set; and (2) obtain an adaptive agent by using this set of strategies as its training opponents.
|
false |
Latent forward model for Real-time Strategy game planning with incomplete information
Real time strategy latent space forward model monte carlo tree search reinforcement learning planning
Model-free deep reinforcement learning approaches have shown superhuman performance in simulated environments (e.g., Atari games, Go, etc). During training, these approaches often implicitly construct a latent space that contains key information for decision making. In this paper, we learn a forward model on this latent space and apply it to model-based planning in miniature Real-time Strategy game with incomplete information (MiniRTS). We first show that the latent space constructed from existing actor-critic models contains relevant information of the game, and design training procedure to learn forward models. We also show that our learned forward model can predict meaningful future state and is usable for latent space Monte-Carlo Tree Search (MCTS), in terms of win rates against rule-based agents.
|
true |
Comparing and Contrasting Deep Learning Weather Prediction Backbones on Navier-Stokes Dynamics
Deep Learning Weather Prediction Navier-Stokes dynamics Benchmark Neural Operators
There has been remarkable progress in the development of Deep Learning Weather Prediction (DLWP) models, so much so that they are poised to become competitive with traditional numerical weather prediction (NWP) models. Indeed, a wide number of DLWP architectures---based on various backbones, including U-Net, Transformer, Graph Neural Network (GNN), or Fourier Neural Operator (FNO)---have demonstrated their potential at forecasting atmospheric states. However, due to differences in training protocols, data choices (resolution, selected prognostic variables, or additional forcing inputs), and forecast horizons, it still remains unclear which of these methods and architectures is most suitable for weather forecasting. Here, we back up and provide a detailed empirical analysis, under controlled conditions, comparing and contrasting the most prominent backbones used in DLWP models. This is done by predicting two-dimensional incompressible Navier-Stokes dynamics with different numbers of parameters and Reynolds number values. In terms of accuracy, memory consumption, and runtime, our results illustrate various tradeoffs, and they show favorable performance of FNO, in comparison with Transformer, U-Net, and GNN backbones.
|
false |
Learning to communicate through imagination with model-based deep multi-agent reinforcement learning
imagination deep reinforcement communication agent set human imagination integral component intelligence core utility
The human imagination is an integral component of our intelligence. Furthermore, the core utility of our imagination is deeply coupled with communication. Language, argued to have been developed through complex interaction within growing collective societies serves as an instruction to the imagination, giving us the ability to share abstract mental representations and perform joint spatiotemporal planning. In this paper, we explore communication through imagination with multi-agent reinforcement learning. Specifically, we develop a model-based approach where agents jointly plan through recurrent communication of their respective predictions of the future. Each agent has access to a learned world model capable of producing model rollouts of future states and predicted rewards, conditioned on the actions sampled from the agent's policy. These rollouts are then encoded into messages and used to learn a communication protocol during training via differentiable message passing. We highlight the benefits of our model-based approach, compared to a set of strong baselines, by developing a set of specialised experiments using novel as well as well-known multi-agent environments.
|
false |
Variational inference for diffusion modulated Cox processes
Cox process variational inference stochastic differential equation smoothing posterior density
This paper proposes a stochastic variational inference (SVI) method for computing an approximate posterior path measure of a Cox process. These processes are widely used in natural and physical sciences, engineering and operations research, and represent a non-trivial model of a wide array of phenomena. In our work, we model the stochastic intensity as the solution of a diffusion stochastic differential equation (SDE), and our objective is to infer the posterior, or smoothing, measure over the paths given Poisson process realizations. We first derive a system of stochastic partial differential equations (SPDE) for the pathwise smoothing posterior density function, a non-trivial result, since the standard solution of SPDEs typically involves an It\^o stochastic integral, which is not defined pathwise. Next, we propose an SVI approach to approximating the solution of the system. We parametrize the class of approximate smoothing posteriors using a neural network, derive a lower bound on the evidence of the observed point process sample-path, and optimize the lower bound using stochastic gradient descent (SGD). We demonstrate the efficacy of our method on both synthetic and real-world problems, and demonstrate the advantage of the neural network solution over standard numerical solvers.
|
false |
Improving Hierarchical Adversarial Robustness of Deep Neural Networks
Adversarial Robustness
Do all adversarial examples have the same consequences? An autonomous driving system misclassifying a pedestrian as a car may induce a far more dangerous --and even potentially lethal-- behavior than, for instance, a car as a bus. In order to better tackle this important problematic, we introduce the concept of hierarchical adversarial robustness. Given a dataset whose classes can be grouped into coarse-level labels, we define hierarchical adversarial examples as the ones leading to a misclassification at the coarse level. To improve the resistance of neural networks to hierarchical attacks, we introduce a hierarchical adversarially robust (HAR) network design that decomposes a single classification task into one coarse and multiple fine classification tasks, before being specifically trained by adversarial defense techniques. As an alternative to an end-to-end learning approach, we show that HAR significantly improves the robustness of the network against $\ell_{\infty}$ and $\ell_{2}$bounded hierarchical attacks on CIFAR-100.
|
false |
Learned Threshold Pruning
Efficiency Model Compression Unstructured Pruning Differentiable Pruning
This paper presents a novel differentiable method for unstructured weight pruning of deep neural networks. Our learned-threshold pruning (LTP) method learns per-layer thresholds via gradient descent, unlike conventional methods where they are set as input. Making thresholds trainable also makes LTP computationally efficient, hence scalable to deeper networks. For example, it takes $30$ epochs for LTP to prune ResNet50 on ImageNet by a factor of $9.1$. This is in contrast to other methods that search for per-layer thresholds via a computationally intensive iterative pruning and fine-tuning process. Additionally, with a novel differentiable $L_0$ regularization, LTP is able to operate effectively on architectures with batch-normalization. This is important since $L_1$ and $L_2$ penalties lose their regularizing effect in networks with batch-normalization. Finally, LTP generates a trail of progressively sparser networks from which the desired pruned network can be picked based on sparsity and performance requirements. These features allow LTP to achieve competitive compression rates on ImageNet networks such as AlexNet ($26.4\times$ compression with $79.1\%$ Top-5 accuracy) and ResNet50 ($9.1\times$ compression with $92.0\%$ Top-5 accuracy). We also show that LTP effectively prunes modern \textit{compact} architectures, such as EfficientNet, MobileNetV2 and MixNet.
|
false |
Using Deep Reinforcement Learning to Train and Evaluate Instructional Sequencing Policies for an Intelligent Tutoring System
Deep Reinforcement Learning Intelligent Tutoring Systems Adaptive policy Instructional Sequencing
We present STEP, a novel Deep Reinforcement Learning solution to the problem of learning instructional sequencing. STEP has three components: 1. Simulate the student by fitting a knowledge tracing model to data logged by an intelligent tutoring system. 2. Train instructional sequencing policies by using Proximal Policy Optimization. 3. Evaluate the learned instructional policies by estimating their local and global impact on learning gains. STEP leverages the student model by representing the student’s knowledge state as a probability vector of knowing each skill and using the student’s estimated learning gains as its reward function to evaluate candidate policies. A learned policy represents a mapping from each state to an action that maximizes the reward, i.e. the upward distance to the next state in the multi-dimensional space. We use STEP to discover and evaluate potential improvements to a literacy and numeracy tutor used by hundreds of children in Tanzania.
|
false |
Siamese Capsule Networks
capsule networks face verification siamse networks few-shot learning contrastive loss
Capsule Networks have shown encouraging results on \textit{defacto} benchmark computer vision datasets such as MNIST, CIFAR and smallNORB. Although, they are yet to be tested on tasks where (1) the entities detected inherently have more complex internal representations and (2) there are very few instances per class to learn from and (3) where point-wise classification is not suitable. Hence, this paper carries out experiments on face verification in both controlled and uncontrolled settings that together address these points. In doing so we introduce \textit{Siamese Capsule Networks}, a new variant that can be used for pairwise learning tasks. We find that the model improves over baselines in the few-shot learning setting, suggesting that capsule networks are efficient at learning discriminative representations when given few samples.
We find that \textit{Siamese Capsule Networks} perform well against strong baselines on both pairwise learning datasets when trained using a contrastive loss with $\ell_2$-normalized capsule encoded pose features, yielding best results in the few-shot learning setting where image pairs in the test set contain unseen subjects.
|
false |
Interactions between Representation Learning and Supervision
representation learning problem supervision representations data continual learning data distribution interactions
Representation learning is one of the fundamental problems of machine learning. On its own, this problem can be cast as an unsupervised dimensionality reduction problem. However, representation learning is often also used as an implicit step in supervised learning (SL) or reinforcement learning (RL) problems. In this paper, we study the possible "interference" supervision, commonly provided through a loss function in SL or a reward function in RL, might have on learning representations, through the lens of learning from limited data and continual learning. Particularly, in connectionist networks, we often face the problem of catastrophic interference whereby changes in the data distribution cause networks to fail to remember previously learned information and learning representations can be done without labeled data. A primary running hypothesis is that representations learned using unsupervised learning are more robust to changes in the data distribution as compared to the intermediate representations learned when using supervision because supervision interferes with otherwise "unconstrained" representation learning objectives. To empirically test hypotheses, we perform experiments using a standard dataset for continual learning, permuted MNIST. Additionally, through a heuristic quantifying the amount of change in the data distribution, we verify that the results are statistically significant.
|
false |
What's in the Box? Exploring the Inner Life of Neural Networks with Robust Rules
Neural Networks CNN explaining interpretable Rules black box
We propose a novel method for exploring how neurons within a neural network interact. In particular, we consider activation values of a network for given data, and propose to mine noise-robust rules of the form $X \rightarrow Y$ , where $X$ and $Y$ are sets of neurons in different layers. To ensure we obtain a small and non-redundant set of high quality rules, we formalize the problem in terms of the Minimum Description Length principle, by which we identify the best set of rules as the one that best compresses the activation data. To discover good rule sets, we propose the unsupervised ExplaiNN algorithm. Extensive evaluation shows that our rules give clear insight in how networks perceive the world: they identify shared, resp. class-specific traits, compositionality within the network, as well as locality in convolutional layers. Our rules are easily interpretable, but also super-charge prototyping as they identify which groups of neurons to consider in unison.
|
true |
Adversarially Guided Actor-Critic
actor tasks methods definite success deep reinforcement problems algorithms sample inefficiency complex environments efficient exploration
Despite definite success in deep reinforcement learning problems, actor-critic algorithms are still confronted with sample inefficiency in complex environments, particularly in tasks where efficient exploration is a bottleneck. These methods consider a policy (the actor) and a value function (the critic) whose respective losses are built using different motivations and approaches. This paper introduces a third protagonist: the adversary. While the adversary mimics the actor by minimizing the KL-divergence between their respective action distributions, the actor, in addition to learning to solve the task, tries to differentiate itself from the adversary predictions. This novel objective stimulates the actor to follow strategies that could not have been correctly predicted from previous trajectories, making its behavior innovative in tasks where the reward is extremely rare. Our experimental analysis shows that the resulting Adversarially Guided Actor-Critic (AGAC) algorithm leads to more exhaustive exploration. Notably, AGAC outperforms current state-of-the-art methods on a set of various hard-exploration and procedurally-generated tasks.
|
false |
A self-explanatory method for the black box problem on discrimination part of CNN
Convolution neural network Interpretability performance Markov random field
Recently, for finding inherent causality implied in CNN, the black box problem of its discrimination part, which is composed of all fully connected layers of the CNN, has been studied by different scientific communities. Many methods were proposed, which can extract various interpretable models from the optimal discrimination part based on inputs and outputs of the part for finding the inherent causality implied in the part. However, the inherent causality cannot readily be found. We think that the problem could be solved by shrinking an interpretable distance which can evaluate the degree for the discrimination part to be easily explained by an interpretable model. This paper proposes a lightweight interpretable model, Deep Cognitive Learning Model(DCLM). And then, a game method between the DCLM and the discrimination part is implemented for shrinking the interpretation distance. Finally, the proposed self-explanatory method was evaluated by some contrastive experiments with certain baseline methods on some standard image processing benchmarks. These experiments indicate that the proposed method can effectively find the inherent causality implied in the discrimination part of the CNN without largely reducing its generalization performance. Moreover, the generalization performance of the DCLM also can be improved.
|
false |
HLogformer: A Hierarchical Transformer for Representing Log Data
Anomaly Detection Hierarchical Transformers Efficient Transformers
Transformers have gained widespread acclaim for their versatility in handling diverse data structures, yet their application to log data remains underexplored. Log data, characterized by its hierarchical, dictionary-like structure, poses unique challenges when processed using conventional transformer models. Traditional methods often rely on manually crafted templates for parsing logs, a process that is labor-intensive and lacks generalizability. Additionally, the linear treatment of log sequences by standard transformers neglects the rich, nested relationships within log entries, leading to suboptimal representations and excessive memory usage.
To address these issues, we introduce HLogformer, a novel hierarchical transformer framework specifically designed for log data. HLogformer leverages the hierarchical structure of log entries to significantly reduce memory costs and enhance representation learning. Unlike traditional models that treat log data as flat sequences, our framework processes log entries in a manner that respects their inherent hierarchical organization. This approach ensures comprehensive encoding of both fine-grained details and broader contextual relationships.
Our contributions are threefold: First, HLogformer is the first framework to design a dynamic hierarchical transformer tailored for dictionary-like log data. Second, it dramatically reduces memory costs associated with processing extensive log sequences. Third, comprehensive experiments demonstrate that HLogformer more effectively encodes hierarchical contextual information, proving to be highly effective for downstream tasks such as synthetic anomaly detection and product recommendation.
|
false |
On the Geometry of Deep Bayesian Active Learning
Bayesian active learning geometric interpretation core-set construction model uncertainty ellipsoid.
We present geometric Bayesian active learning by disagreements (GBALD), a framework that performs BALD on its geometric interpretation interacting with a deep learning model. There are two main components in GBALD: initial acquisitions based on core-set construction and model uncertainty estimation with those initial acquisitions. Our key innovation is to construct the core-set on an ellipsoid, not typical sphere, preventing its updates towards the boundary regions of the distributions. Main improvements over BALD are twofold: relieving sensitivity to uninformative prior and reducing redundant information of model uncertainty. To guarantee the improvements, our generalization analysis proves that, compared to typical Bayesian spherical interpretation, geodesic search with ellipsoid can derive a tighter lower error bound and achieve higher probability to obtain a nearly zero error. Experiments on acquisitions with several scenarios demonstrate that, yielding slight perturbations to noisy and repeated samples, GBALD further achieves significant accuracy improvements than BALD, BatchBALD and other baselines.
|
true |
Are wider nets better given the same number of parameters?
network width over-parametrization understanding deep learning
Empirical studies demonstrate that the performance of neural networks improves with increasing number of parameters. In most of these studies, the number of parameters is increased by increasing the network width. This begs the question: Is the observed improvement due to the larger number of parameters, or is it due to the larger width itself? We compare different ways of increasing model width while keeping the number of parameters constant. We show that for models initialized with a random, static sparsity pattern in the weight tensors, network width is the determining factor for good performance, while the number of weights is secondary, as long as the model achieves high training accuarcy. As a step towards understanding this effect, we analyze these models in the framework of Gaussian Process kernels. We find that the distance between the sparse finite-width model kernel and the infinite-width kernel at initialization is indicative of model performance.
|
false |
Multi-scale Network Architecture Search for Object Detection
Object Detection Neural Architecture Search
Many commonly-used detection frameworks aim to handle the multi-scale object detection problem. The input image is always encoded to multi-scale features and objects grouped by scale range are assigned to the corresponding features. However, the design of multi-scale feature production is quite hand-crafted or partially automatic. In this paper, we show that more possible architectures of encoder network and different strategies of feature utilization can lead to superior performance. Specifically, we propose an efficient and effective multi-scale network architecture search method (MSNAS) to improve multi-scale object detection by jointly optimizing network stride search of the encoder and appropriate feature selection for detection heads. We demonstrate the effectiveness of the method on COCO dataset and obtain a remarkable performance gain with respect to the original Feature Pyramid Networks.
|
true |
Guided Autoregressive Diffusion Models with Applications to PDE Simulation
neural PDE solver PDE partial differential equation forecasting data-assimilation diffusion denoising autoregressive neural surrogate reconstruction guidance
Solving partial differential equations (PDEs) is of crucial importance in science and engineering. Yet numerical solvers necessitate high space-time resolution which in turn leads to heavy computational cost. Often applications require solving the same PDE many times, only changing initial conditions or parameters. In this setting, data-driven machine learning methods have shown great promise, a principle advantage being the ability to simultaneously train at coarse resolutions and produce fast PDE solutions. In this work we introduce the Guided AutoRegressive Diffusion model (GuARD), which is trained over short segments from PDE trajectories and a posteriori sampled by conditioning over (1) some initial state to tackle forecasting and/or over (2) some sparse space-time observations for data assimilation purposes. We empirically demonstrate the ability of such a sampling procedure to generate accurate predictions of long PDE trajectories.
|
true |
Rethinking Architecture Selection in Differentiable NAS
architecture selection supernet architecture parameters much optimization operation darts differentiable nas
Differentiable Neural Architecture Search is one of the most popular Neural Architecture Search (NAS) methods for its search efficiency and simplicity, accomplished by jointly optimizing the model weight and architecture parameters in a weight-sharing supernet via gradient-based algorithms. At the end of the search phase, the operations with the largest architecture parameters will be selected to form the final architecture, with the implicit assumption that the values of architecture parameters reflect the operation strength. While much has been discussed about the supernet's optimization, the architecture selection process has received little attention. We provide empirical and theoretical analysis to show that the magnitude of architecture parameters does not necessarily indicate how much the operation contributes to the supernet's performance. We propose an alternative perturbation-based architecture selection that directly measures each operation's influence on the supernet. We re-evaluate several differentiable NAS methods with the proposed architecture selection and find that it is able to extract significantly improved architectures from the underlying supernets consistently. Furthermore, we find that several failure modes of DARTS can be greatly alleviated with the proposed selection method, indicating that much of the poor generalization observed in DARTS can be attributed to the failure of magnitude-based architecture selection rather than entirely the optimization of its supernet.
|
false |
ACT: Asymptotic Conditional Transport
Statistical distance Divergence Optimal Transport Implicit Distribution Deep Generative Models GANs
We propose conditional transport (CT) as a new divergence to measure the difference between two probability distributions. The CT divergence consists of the expected cost of a forward CT, which constructs a navigator to stochastically transport a data point of one distribution to the other distribution, and that of a backward CT which reverses the transport direction. To apply it to the distributions whose probability density functions are unknown but random samples are accessible, we further introduce asymptotic CT (ACT), whose estimation only requires access to mini-batch based discrete empirical distributions. Equipped with two navigators that amortize the computation of conditional transport plans, the ACT divergence comes with unbiased sample gradients that are straightforward to compute, making it amenable to mini-batch stochastic gradient descent based optimization. When applied to train a generative model, the ACT divergence is shown to strike a good balance between mode covering and seeking behaviors and strongly resist mode collapse. To model high-dimensional data, we show that it is sufficient to modify the adversarial game of an existing generative adversarial network (GAN) to a game played by a generator, a forward navigator, and a backward navigator, which try to minimize a distribution-to-distribution transport cost by optimizing both the distribution of the generator and conditional transport plans specified by the navigators, versus a critic that does the opposite by inflating the point-to-point transport cost. On a wide variety of benchmark datasets for generative modeling, substituting the default statistical distance of an existing GAN with the ACT divergence is shown to consistently improve the performance.
|
true |
How Does Mixup Help With Robustness and Generalization?
Mixup adversarial robustness generalization
Mixup is a popular data augmentation technique based on on convex combinations of pairs of examples and their labels. This simple technique has shown to substantially improve both the model's robustness as well as the generalization of the trained model. However, it is not well-understood why such improvement occurs. In this paper, we provide theoretical analysis to demonstrate how using Mixup in training helps model robustness and generalization. For robustness, we show that minimizing the Mixup loss corresponds to approximately minimizing an upper bound of the adversarial loss. This explains why models obtained by Mixup training exhibits robustness to several kinds of adversarial attacks such as Fast Gradient Sign Method (FGSM). For generalization, we prove that Mixup augmentation corresponds to a specific type of data-adaptive regularization which reduces overfitting. Our analysis provides new insights and a framework to understand Mixup.
|
true |
Passage Ranking with Weak Supervision
Passage Ranking Weak Supervision BERT Models
In this paper, we propose a \textit{weak supervision} framework for neural ranking tasks based on the data programming paradigm \citep{Ratner2016}, which enables us to leverage multiple weak supervision signals from different sources. Empirically, we consider two sources of weak supervision signals, unsupervised ranking functions and semantic feature similarities. We train a BERT-based passage-ranking model (which achieves new state-of-the-art performances on two benchmark datasets with full supervision) in our weak supervision framework. Without using ground-truth training labels, BERT-PR models outperform BM25 baseline by a large margin on all three datasets and even beat the previous state-of-the-art results with full supervision on two of datasets.
|
false |
Counterfactual Fairness through Data Preprocessing
Counterfactual fairness data preprocessing fairness test discrimination detection affirmative action
Machine learning has become more important in real-life decision-making but people are concerned about the ethical problems it may bring when used improperly. Recent work brings the discussion of machine learning fairness into the causal framework and elaborates on the concept of Counterfactual Fairness. In this paper, we develop the Fair Learning through dAta Preprocessing (FLAP) algorithm to learn counterfactually fair decisions from biased training data and formalize the conditions where different data preprocessing procedures should be used to guarantee counterfactual fairness. We also show that Counterfactual Fairness is equivalent to the conditional independence of the decisions and the sensitive attributes given the processed non-sensitive attributes, which enables us to detect discrimination in the original decision using the processed data. The performance of our algorithm is illustrated using simulated data and real-world applications.
|
true |
DISENTANGLED STATE SPACE MODELS: UNSUPERVISED LEARNING OF DYNAMICS ACROSS HETEROGENEOUS ENVIRONMENTS
State Space Models Sequential Data Bayesian Filtering Amortized Variational Inference Disentangled Representations Video Analysis
Sequential data often originates from diverse environments. Across them exist both shared regularities and environment specifics. To learn robust cross-environment descriptions of sequences we introduce disentangled state space models (DSSM). In the latent space of DSSM environment-invariant state dynamics is explicitly disentangled from environment-specific information governing that dynamics. We empirically show that such separation enables robust prediction, sequence manipulation and environment characterization. We also propose an unsupervised VAE-based training procedure to learn DSSM as Bayesian filters. In our experiments, we demonstrate state-of-the-art performance in controlled generation and prediction of bouncing ball video sequences across varying gravitational influences.
|
false |
Maximum Categorical Cross Entropy (MCCE): A noise-robust alternative loss function to mitigate racial bias in Convolutional Neural Networks (CNNs) by reducing overfitting
models mcce convolutional neural networks cnns cce colorferet alternative loss function racial bias cnn models
Categorical Cross Entropy (CCE) is the most commonly used loss function in deep neural networks such as Convolutional Neural Networks (CNNs) for multi-class classification problems. In spite of the fact that CCE is highly susceptible to noise; CNN models trained without accounting for the unique noise characteristics of the input data, or noise introduced during model training, invariably suffer from overfitting affecting model generalizability. The lack of generalizability becomes especially apparent in the context of ethnicity/racial image classification problems encountered in the domain of computer vision. One such problem is the unintended discriminatory racial bias that CNN models trained using CCE fail to adequately address. In other words, CNN models trained using CCE offer a skewed representation of classification performance favoring lighter skin tones.
In this paper, we propose and empirically validate a novel noise-robust extension to the existing CCE loss function called Maximum Categorical Cross-Entropy (MCCE), which utilizes CCE loss and a novel reconstruction loss, calculated using the Maximum Entropy (ME) measures of the convolutional kernel weights and input training dataset. We compare the use of MCCE with CCE-trained models on two benchmarking datasets, colorFERET and UTKFace, using a Residual Network (ResNet) CNN architecture. MCCE-trained models reduce overfitting by 5.85% and 4.3% on colorFERET and UTKFace datasets respectively. In cross-validation testing, MCCE-trained models outperform CCE-trained models by 8.8% and 25.16% on the colorFERET and UTKFace datasets respectively. MCCE addresses and mitigates the persistent problem of inadvertent racial bias for facial recognition problems in the domain of computer vision.
|
false |
Modifying Memories in Transformer Models
Transformers memorization question answering
Large Transformer models have achieved impressive performance in many natural language tasks. In particular, Transformer based language models have been shown to have great capabilities in encoding factual knowledge in their vast amount of parameters. While the tasks of improving the memorization and generalization of Transformers have been widely studied, it is not well known how to make transformers forget specific old facts and memorize new ones. In this paper, we propose a new task of explicitly modifying specific factual knowledge in Transformer models while ensuring the model performance does not degrade on the unmodified facts. This task is useful in many scenarios, such as updating stale knowledge, protecting privacy, and eliminating unintended biases stored in the models. We benchmarked several approaches that provide natural baseline performances on this task. This leads to the discovery of key components of a Transformer model that are especially effective for knowledge modifications. The work also provides insights into the role that different training phases (such as pretraining and fine-tuning) play towards memorization and knowledge modification.
|
false |
Safety Verification of Model Based Reinforcement Learning Controllers
Reachable set state constraints safety verification model-based reinforcement learning
Model-based reinforcement learning (RL) has emerged as a promising tool for developing controllers for real world systems (e.g., robotics, autonomous driving, etc.). However, real systems often have constraints imposed on their state space which must be satisfied to ensure the safety of the system and its environment. Developing a verification tool for RL algorithms is challenging because the non-linear structure of neural networks impedes analytical verification of such models or controllers. To this end, we present a novel safety verification framework for model-based RL controllers using reachable set analysis. The proposed framework can efficiently handle models and controllers which are represented using neural networks. Additionally, if a controller fails to satisfy the safety constraints in general, the proposed framework can also be used to identify the subset of initial states from which the controller can be safely executed.
|
true |
HEP-JEPA: A foundation model for collider physics
foundation model self-supervised learning particle physics high energy physics jepa
We present a transformer architecture-based foundation model for tasks at high-energy particle colliders such as the Large Hadron Collider. We train the model to classify jets using a self-supervised strategy inspired by the Joint Embedding Predictive Architecture. We use the JetClass dataset containing 100M jets of various known particles to pre-train the model with a data-centric approach --- the model uses a fraction of the jet constituents as the context to predict the embeddings of the unseen target constituents. Our pre-trained model fares well with other datasets for standard classification benchmark tasks. We test our model on two additional downstream tasks: top tagging and differentiating light-quark jets from gluon jets. We also evaluate our model with task-specific metrics and baselines and compare it with state-of-the-art models in high-energy physics. Therefore, this work contributes to the development of scientific foundation models by demonstrating how self-supervised transformer architectures can extract deep insights from high-energy physics data.
|
true |
Clairvoyance: A Pipeline Toolkit for Medical Time Series
reproducibility healthcare medical time series pipeline toolkit software
Time-series learning is the bread and butter of data-driven *clinical decision support*, and the recent explosion in ML research has demonstrated great potential in various healthcare settings. At the same time, medical time-series problems in the wild are challenging due to their highly *composite* nature: They entail design choices and interactions among components that preprocess data, impute missing values, select features, issue predictions, estimate uncertainty, and interpret models. Despite exponential growth in electronic patient data, there is a remarkable gap between the potential and realized utilization of ML for clinical research and decision support. In particular, orchestrating a real-world project lifecycle poses challenges in engineering (i.e. hard to build), evaluation (i.e. hard to assess), and efficiency (i.e. hard to optimize). Designed to address these issues simultaneously, Clairvoyance proposes a unified, end-to-end, autoML-friendly pipeline that serves as a (i) software toolkit, (ii) empirical standard, and (iii) interface for optimization. Our ultimate goal lies in facilitating transparent and reproducible experimentation with complex inference workflows, providing integrated pathways for (1) personalized prediction, (2) treatment-effect estimation, and (3) information acquisition. Through illustrative examples on real-world data in outpatient, general wards, and intensive-care settings, we illustrate the applicability of the pipeline paradigm on core tasks in the healthcare journey. To the best of our knowledge, Clairvoyance is the first to demonstrate viability of a comprehensive and automatable pipeline for clinical time-series ML.
|
false |
ORTHOGONAL SAE: FEATURE DISENTANGLEMENT THROUGH COMPETITION-AWARE ORTHOGONALITY CONSTRAINTS
Feature Disentanglement Debiasing in Neural Networks Dangerous Knowledge Filtering Ethical AI Sparse Autoencoders Model Safety Bias Reduction Techniques
Understanding the internal representations of large language models is crucial for ensuring their reliability and enabling targeted interventions, with sparse autoencoders (SAEs) emerging as a promising approach for decomposing neural activations into interpretable features. A key challenge in SAE development is feature absorption, where features stop firing independently and are ``absorbed'' into each other to minimize $L_1$ penalty. We address this through Orthogonal SAE, which introduces sparsity-guided orthogonality constraints that dynamically identify and disentangle competing features through a principled three-phase curriculum. Our approach achieves state-of-the-art results on the Gemma-2-2B language model for feature absorption while maintaining strong reconstruction quality and model preservation on downstream tasks. These results demonstrate that orthogonality constraints and competition-aware training can effectively balance the competing objectives of feature interpretability and model fidelity, enabling more reliable analysis of neural network representations.
|
false |
Efficient estimates of optimal transport via low-dimensional embeddings
optimal transport sinkhorn divergences robustness neural networks lipschitz spectral norm
Optimal transport distances (OT) have been widely used in recent work in Machine Learning as ways to compare probability distributions. These are costly to compute when the data lives in high dimension.
Recent work aims specifically at reducing this cost by computing OT using low-rank projections of the data (seen as discrete measures)~\citep{paty2019subspace}. We extend this approach and show that one can approximate OT distances by using more general families of maps provided they are 1-Lipschitz. The best estimate is obtained by maximising OT over the given family. As OT calculations are done after mapping data to a lower dimensional space, our method scales well with the original data dimension.
We demonstrate the idea with neural networks.
|
false |
ChunkRAG: A Novel LLM-Chunk Filtering Method for RAG Systems
LLM RAG Chunking Fact Checking Retrieval Information Retrieval
Retrieval-Augmented Generation (RAG) frameworks leveraging large language models (LLMs) frequently retrieve extraneous or weakly relevant information, leading to factual inaccuracies and hallucinations in generated responses. Existing document-level retrieval approaches lack sufficient granularity to effectively filter non-essential content. This paper introduces ChunkRAG, a retrieval framework that refines information selection through semantic chunking and chunk-level evaluation. ChunkRAG applies a dynamic greedy chunk aggregation strategy to segment documents into semantically coherent, variable-length sections based on cosine similarity. Empirical evaluations on the PopQA, PubHealth and Biography dataset indicate that ChunkRAG improves response accuracy over state-of-the-art RAG methods. The analysis further demonstrates that chunk-level filtering reduces redundant and weakly related information, enhancing the factual consistency of responses. By incorporating fine-grained retrieval mechanisms, ChunkRAG provides a scalable and domain-agnostic approach to mitigate hallucinations in knowledge-intensive tasks such as fact-checking and multi-hop reasoning.
|
true |
Understanding Few-Shot Multi-Task Representation Learning Theory
multi-task learning few-shot learning learning theory
Multi-Task Representation Learning (MTR) is a popular paradigm to learn shared representations from multiple related tasks. It has demonstrated its efficiency for solving different problems, ranging from machine translation for natural language processing to object detection in computer vision. On the other hand, Few-Shot Learning is a recent problem that seeks to mimic the human capability to quickly learn how to solve a target task with little supervision. For this topic, researchers have turned to meta-learning that learns to learn a new task by training a model on a lot of small tasks. As meta-learning still suffers from a lack of theoretical understanding for its success in few-shot tasks, an intuitively appealing approach would be to bridge the gap between it and multi-task learning to better understand the former using the results established for the latter. In this post, we dive into a recent ICLR 2021 paper by Du et al. that demonstrated novel learning bounds for multi-task learning in the few-shot setting and go beyond it by establishing the connections that allow to better understand the inner workings of meta-learning algorithms as well.
|
false |
Robust Multi-Agent Reinforcement Learning Driven by Correlated Equilibrium
Robust Multi-agent Reinforcement Learning Correlated Equilibrium
In this paper we deal with robust cooperative multi-agent reinforcement learning (CMARL). While CMARL has many potential applications, only a trained policy that is robust enough can be confidently deployed in real world. Existing works on robust MARL mainly apply vanilla adversarial training in centralized training and decentralized execution paradigm. We, however, find that if a CMARL environment contains an adversarial agent, the performance of decentralized equilibrium might perform significantly poor for achieving such adversarial robustness. To tackle this issue, we suggest that when execution the non-adversarial agents must jointly make the decision to improve the robustness, therefore solving correlated equilibrium instead. We theoretically demonstrate the superiority of correlated equilibrium over the decentralized one in adversarial MARL settings. Therefore, to achieve robust CMARL, we introduce novel strategies to encourage agents to learn correlated equilibrium while maximally preserving the convenience of the decentralized execution. The global variables with mutual information are proposed to help agents learn robust policies with MARL algorithms. The experimental results show that our method can dramatically boost performance on the SMAC environments.
|
true |
SkipW: Resource Adaptable RNN with Strict Upper Computational Limit
Recurrent neural networks Flexibility Computational resources
We introduce Skip-Window, a method to allow recurrent neural networks (RNNs) to trade off accuracy for computational cost during the analysis of a sequence. Similarly to existing approaches, Skip-Window extends existing RNN cells by adding a mechanism to encourage the model to process fewer inputs. Unlike existing approaches, Skip-Window is able to respect a strict computational budget, making this model more suitable for limited hardware. We evaluate this approach on two datasets: a human activity recognition task and adding task. Our results show that Skip-Window is able to exceed the accuracy of existing approaches for a lower computational cost while strictly limiting said cost.
|
true |
Systematic generalisation with group invariant predictions
Systematic generalisation invariance penalty semantic anomaly detection
We consider situations where the presence of dominant simpler correlations with the target variable in a training set can cause an SGD-trained neural network to be less reliant on more persistently correlating complex features. When the non-persistent, simpler correlations correspond to non-semantic background factors, a neural network trained on this data can exhibit dramatic failure upon encountering systematic distributional shift, where the correlating background features are recombined with different objects. We perform an empirical study on three synthetic datasets, showing that group invariance methods across inferred partitionings of the training set can lead to significant improvements at such test-time situations. We also suggest a simple invariance penalty, showing with experiments on our setups that it can perform better than alternatives. We find that even without assuming access to any systematically shifted validation sets, one can still find improvements over an ERM-trained reference model.
|
true |
You Only Need Adversarial Supervision for Semantic Image Synthesis
Semantic Image Synthesis GANs Image Generation Deep Learning
Despite their recent successes, GAN models for semantic image synthesis still suffer from poor image quality when trained with only adversarial supervision. Historically, additionally employing the VGG-based perceptual loss has helped to overcome this issue, significantly improving the synthesis quality, but at the same time limiting the progress of GAN models for semantic image synthesis. In this work, we propose a novel, simplified GAN model, which needs only adversarial supervision to achieve high quality results. We re-design the discriminator as a semantic segmentation network, directly using the given semantic label maps as the ground truth for training. By providing stronger supervision to the discriminator as well as to the generator through spatially- and semantically-aware discriminator feedback, we are able to synthesize images of higher fidelity with better alignment to their input label maps, making the use of the perceptual loss superfluous. Moreover, we enable high-quality multi-modal image synthesis through global and local sampling of a 3D noise tensor injected into the generator, which allows complete or partial image change. We show that images synthesized by our model are more diverse and follow the color and texture distributions of real images more closely. We achieve an average improvement of $6$ FID and $5$ mIoU points over the state of the art across different datasets using only adversarial supervision.
|
false |
Localized Meta-Learning: A PAC-Bayes Analysis for Meta-Learning Beyond Global Prior
localized meta-learning PAC-Bayes meta-learning
Meta-learning methods learn the meta-knowledge among various training tasks and aim to promote the learning of new tasks under the task similarity assumption. Such meta-knowledge is often represented as a fixed distribution; this, however, may be too restrictive to capture various specific task information because the discriminative patterns in the data may change dramatically across tasks. In this work, we aim to equip the meta learner with the ability to model and produce task-specific meta-knowledge and, accordingly, present a localized meta-learning framework based on the PAC-Bayes theory. In particular, we propose a Local Coordinate Coding (LCC) based prior predictor that allows the meta learner to generate local meta-knowledge for specific tasks adaptively. We further develop a practical algorithm with deep neural network based on the bound. Empirical results on real-world datasets demonstrate the efficacy of the proposed method.
|
true |
Object Representations as Fixed Points: Training Iterative Inference Algorithms with Implicit Differentiation
implicit differentiation object-centric learning iterative amortized inference symmetric generative models
Deep generative models, particularly those that aim to factorize the observations into discrete entities (such as objects), must often use iterative inference procedures that break symmetries among equally plausible explanations for the data. Such inference procedures include variants of the expectation-maximization algorithm and structurally resemble clustering algorithms in a latent space. However, combining such methods with deep neural networks necessitates differentiating through the inference process, which can make optimization exceptionally challenging. In this work, we observe that such iterative inference methods can be made differentiable by means of the implicit function theorem, and develop an implicit differentiation approach that improves the stability and tractability of training such models by decoupling the forward and backward passes. This connection enables us to apply recent advances in optimizing implicit layers to not only improve the stability and optimization of the slot attention module in SLATE, a state-of-the-art method for learning entity representations, but do so with constant space and time complexity in backpropagation and only one additional line of code.
|
true |
ReMixer: Object-aware Mixing Layer for Vision Transformers and Mixers
vision transformers patch-based models object-centric learning
Patch-based models, e.g., Vision Transformers (ViTs) and Mixers, have shown impressive results on various visual recognition tasks, exceeding classic convolutional networks. While the initial patch-based models treated all patches equally, recent studies reveal that incorporating inductive biases like spatiality benefits the learned representations. However, most prior works solely focused on the position of patches, overlooking the scene structure of images. This paper aims to further guide the interaction of patches using the object information. Specifically, we propose ReMixer, which reweights the patch mixing layers based on the patch-wise object labels extracted from pretrained saliency or classification models. We apply ReMixer on various patch-based models using different patch mixing layers: ViT, MLP-Mixer, and ConvMixer, where our method consistently improves the classification accuracy and background robustness of baseline models.
|
true |
Which Language Evolves Between Heterogeneous Agents? - Communicating Movement Instructions With Widely Different Time Scopes
Emergent Communication Dreaming Reinforcement Learning Plan Execution Topographic Similarity
This paper studies the evolving communication between two agents, a listener and speaker, in a plan execution task in which the speaker needs to communicate the plan to the acting agent, while operating on different time scales.
We analyse the topographic similarity of the resulting language learned by the proposed imagination-based learning process.
As the speaker agent perceives the movement space strictly in absolute coordinates and the actor can only choose relative actions in the movement space, we can show that the structure of their emergent communication is not predestined.
Both relative and absolute encodings of desired movements can develop by chance in this setting, but we can alter the chance by using a population of learners.
We conclude that our imagination-based learning strategy successfully breaks the strict hierarchy between planner and executioner.
|
true |
Grounding Physical Concepts of Objects and Events Through Dynamic Visual Reasoning
Concept Learning Neuro-Symbolic Learning Video Reasoning Visual Reasoning
We study the problem of dynamic visual reasoning on raw videos. This is a challenging problem; currently, state-of-the-art models often require dense supervision on physical object properties and events from simulation, which are impractical to obtain in real life. In this paper, we present the Dynamic Concept Learner (DCL), a unified framework that grounds physical objects and events from video and language. DCL first adopts a trajectory extractor to track each object over time and to represent it as a latent, object-centric feature vector. Building upon this object-centric representation, DCL learns to approximate the dynamic interaction among objects using graph networks. DCL further incorporates a semantic parser to parse question into semantic programs and, finally, a program executor to run the program to answer the question, levering the learned dynamics model. After training, DCL can detect and associate objects across the frames, ground visual properties and physical events, understand the causal relationship between events, make future and counterfactual predictions, and leverage these extracted presentations for answering queries. DCL achieves state-of-the-art performance on CLEVRER, a challenging causal video reasoning dataset, even without using ground-truth attributes and collision labels from simulations for training. We further test DCL on a newly proposed video-retrieval and event localization dataset derived from CLEVRER, showing its strong generalization capacity.
|
true |
Recent Advances in Deep Learning for Routing Problems
Deep Learning Travelling Salesperson Problem Routing Problems Combinatorial Optimization Graph Neural Networks
Developing neural network-driven solvers for combinatorial optimization problems such as the Travelling Salesperson Problem have seen a surge of academic interest recently. This blogpost presents a Neural Combinatorial Optimization pipeline that unifies several recently proposed model architectures and learning paradigms into one single framework. Through the lens of the pipeline, we analyze recent advances in deep learning for routing problems, and provide new directions to stimulate future research towards practical impact.
|
false |
Mutual Calibration between Explicit and Implicit Deep Generative Models
deep generative models generative adversarial networks density estimation
Deep generative models are generally categorized into explicit models and implicit models. The former defines an explicit density form that allows likelihood inference; while the latter targets a flexible transformation from random noise to generated samples. To take full advantages of both models, we propose Stein Bridging, a novel joint training framework that connects an explicit (unnormalized) density estimator and an implicit sample generator via Stein discrepancy. We show that the Stein bridge 1) induces novel mutual regularization via kernel Sobolev norm penalization and Moreau-Yosida regularization, and 2) stabilizes the training dynamics. Empirically, we demonstrate that Stein Bridging can facilitate the density estimator to accurately identify data modes and guide the sample generator to output more high-quality samples especially when the training samples are contaminated or limited.
|
true |
Decoupling Global and Local Representations via Invertible Generative Flows
Generative Models Generative Flow Normalizing Flow Image Generation Representation Learning
In this work, we propose a new generative model that is capable of automatically decoupling global and local representations of images in an entirely unsupervised setting, by embedding a generative flow in the VAE framework to model the decoder.
Specifically, the proposed model utilizes the variational auto-encoding framework to learn a (low-dimensional) vector of latent variables to capture the global information of an image, which is fed as a conditional input to a flow-based invertible decoder with architecture borrowed from style transfer literature.
Experimental results on standard image benchmarks demonstrate the effectiveness of our model in terms of density estimation, image generation and unsupervised representation learning.
Importantly, this work demonstrates that with only architectural inductive biases, a generative model with a likelihood-based objective is capable of learning decoupled representations, requiring no explicit supervision.
The code for our model is available at \url{https://github.com/XuezheMax/wolf}.
|
false |
Model-Free Counterfactual Credit Assignment
credit assignment model-free RL causality hindsight
Credit assignment in reinforcement learning is the problem of measuring an action’s influence on future rewards.
In particular, this requires separating \emph{skill} from \emph{luck}, ie.\ disentangling the effect of an action on rewards from that of external factors and subsequent actions. To achieve this, we adapt the notion of counterfactuals from causality theory to a model-free RL setup.
The key idea is to condition value functions on \emph{future} events, by learning to extract relevant information from a trajectory. We then propose to use these as future-conditional baselines and critics in policy gradient algorithms and we develop a valid, practical variant with provably lower variance, while achieving unbiasedness by constraining the hindsight information not to contain information about the agent’s actions. We demonstrate the efficacy and validity of our algorithm on a number of illustrative problems.
|
false |
TOMA: Topological Map Abstraction for Reinforcement Learning
Planning Reinforcement Learning Representation Learning
Animals are able to discover the topological map (graph) of surrounding environment, which will be used for navigation. Inspired by this biological phenomenon, researchers have recently proposed to learn a graph representation for Markov decision process (MDP) and use such graphs for planning in reinforcement learning (RL). However, existing learning-based graph generation methods suffer from many drawbacks. One drawback is that existing methods do not learn an abstraction for graphs, which results in high memory and computation cost. This drawback also makes generated graph non-robust, which degrades the planning performance. Another drawback is that existing methods cannot be used for facilitating exploration which is important in RL. In this paper, we propose a new method, called topological map abstraction (TOMA), for graph generation. TOMA can learn an abstract graph representation for MDP, which costs much less memory and computation cost than existing methods. Furthermore, TOMA can be used for facilitating exploration. In particular, we propose planning to explore, in which TOMA is used to accelerate exploration by guiding the agent towards unexplored states. A novel experience replay module called vertex memory is also proposed to improve exploration performance. Experimental results show that TOMA can outperform existing methods to achieve the state-of-the-art performance.
|
true |
Simulate Time-integrated Coarse-grained Molecular Dynamics with Geometric Machine Learning
molecular dynamics learning to simulate coarse-graining graph neural network score-based generative models polymer material science
Molecular dynamics (MD) simulation is the workhorse of various scientific domains but is limited by high computational cost. Learning-based force fields have made major progress in accelerating ab-initio MD simulation but are still not fast enough for many real-world applications that require long-time MD simulation. In this paper, we adopt a different machine learning approach where we coarse-grain a physical system using graph clustering and model the system evolution with a very large time-integration step using graph neural networks. A novel score-based GNN refinement module resolves the long-standing challenge of long-time simulation instability. Despite only being trained with short MD trajectory data, our learned simulator can generalize to unseen novel systems, and simulate for much longer than the training trajectories. Properties requiring 10-100 ns level long-time dynamics can be accurately recovered at several orders of magnitude higher speed than classical force fields. We demonstrate the effectiveness of our method on two realistic complex systems: (1) single-chain coarse-grained polymers in implicit solvent; (2) multi-component Li-ion polymer electrolyte systems.
|
true |
Improved Adversarial Image Captioning
image captioning discrete GAN training
In this paper we study image captioning as a conditional GAN training, proposing both a context-aware LSTM captioner and co-attentive discriminator, which enforces semantic alignment between images and captions. We investigate the viability of two discrete GAN training methods: Self-critical Sequence Training (SCST) and Gumbel Straight-Through (ST) and demonstrate that SCST shows more stable gradient behavior and improved results over Gumbel ST.
|
true |
Large Language Models powered Neural Solvers for Generalized Vehicle Routing Problems
Large Language Models Neural Combinatorial Optimization Vehicle Routing Problems
Neural Combinatorial Optimization (NCO) has shown promise in solving combinatorial optimization problems end-to-end with minimal expert-driven algorithm design. However, existing constructive NCO methods for Vehicle Routing Problems (VRPs) often rely on attention-based node selection mechanisms that struggle with large-scale instances.
To address this, we propose a directed fine-tuning approach for NCO based on LLM-driven automatic heuristic design. We first introduce an evolution-driven process that extracts implicit structural features from input instances, forming LLM-guided attention bias. This bias is then integrated into the neural model’s attention scores, enhancing solution flexibility and scalability. Instead of retraining from scratch, we fine-tune the model on a small, diverse dataset to transfer learned heuristics effectively to larger problem instances.
Experimental results show that our approach achieves state-of-the-art performance on TSP and CVRP, significantly improving generalization to both synthetic and real-world datasets (TSPLIB and CVRPLIB) with thousands of nodes.
|
true |
Iterated learning for emergent systematicity in VQA
iterated learning cultural transmission neural module network clevr shapes vqa visual question answering systematic generalization compositionality
Although neural module networks have an architectural bias towards compositionality, they require gold standard layouts to generalize systematically in practice. When instead learning layouts and modules jointly, compositionality does not arise automatically and an explicit pressure is necessary for the emergence of layouts exhibiting the right structure. We propose to address this problem using iterated learning, a cognitive science theory of the emergence of compositional languages in nature that has primarily been applied to simple referential games in machine learning. Considering the layouts of module networks as samples from an emergent language, we use iterated learning to encourage the development of structure within this language. We show that the resulting layouts support systematic generalization in neural agents solving the more complex task of visual question-answering. Our regularized iterated learning method can outperform baselines without iterated learning on SHAPES-SyGeT (SHAPES Systematic Generalization Test), a new split of the SHAPES dataset we introduce to evaluate systematic generalization, and on CLOSURE, an extension of CLEVR also designed to test systematic generalization. We demonstrate superior performance in recovering ground-truth compositional program structure with limited supervision on both SHAPES-SyGeT and CLEVR.
|
false |
Generative Adversarial User Privacy in Lossy Single-Server Information Retrieval
Generative adversarial network generative adversarial privacy information-theoretic privacy compression private information retrieval data-driven framework
We consider the problem of information retrieval from a dataset of files stored on a single server under both a user distortion and a user privacy constraint. Specifically, a user requesting a file from the dataset should be able to reconstruct the requested file with a prescribed distortion, and in addition, the identity of the requested file should be kept private from the server with a prescribed privacy level. The proposed model can be seen as an extension of the well-known concept of private information retrieval by allowing for distortion in the retrieval process and relaxing the perfect privacy requirement. We initiate the study of the tradeoff between download rate, distortion, and user privacy leakage, and show that the optimal rate-distortion-leakage tradeoff is convex and that it allows for a concise information-theoretical formulation in terms of mutual information in the limit of large file sizes. Moreover, we propose a new data-driven framework by leveraging recent advancements in generative adversarial models which allows a user to learn efficient schemes in terms of download rate from the data itself. Learning the scheme is formulated as a constrained minimax game between a user which desires to keep the identity of the requested file private and an adversary that tries to infer which file the user is interested in under a distortion constraint. In general, guaranteeing a certain privacy level leads to a higher rate-distortion tradeoff curve, and hence a sacrifice in either download rate or distortion. We evaluate the performance of the scheme on a synthetic Gaussian dataset as well as on the MNIST dataset. For the MNIST dataset, the data-driven approach significantly outperforms a proposed general achievable scheme combining source coding with the download of multiple files.
|
true |
Why Are Web AI Agents More Vulnerable Than Standalone LLMs? A Security Analysis
AI Agent Web AI Agent LLM Jailbreaking
Recent research has significantly advanced Web AI agents, introducing groundbreaking architectures and benchmarks demonstrating major progress in autonomous web interaction and navigation. However, recent studies have shown that many AI agents can execute malicious tasks and are more vulnerable than standalone LLMs. Our work studies why Web AI agents, built on safety-aligned backbone Large Language Models (LLMs), remain highly susceptible to following malicious user inputs. In particular, we investigate the sources of these vulnerabilities by analyzing the differences between Web AI agents and standalone LLMs in terms of their design and components, quantifying the vulnerability rate introduced by each component. Through a fine-grained evaluation to uncover nuanced jailbreaking signals, we identify three key factors in Web AI agents that make them more vulnerable than standalone LLMs: 1) directly including user input in the system prompt of LLMs, 2) generating actions in a multi-step manner, and 3) processing Event Streams (observation + action history) from web navigation. Furthermore, we observe that many current benchmarks and evlautions rely on mock-up websites, which could potentially lead to misleading results. Our findings highlight the need to prioritize security and robustness when designing the individual components of AI agents. We also suggest developing more realistic safety evaluation systems for Web AI agents.
|
true |
Implicit Normalizing Flows
Normalizing flows deep generative models probabilistic inference implicit functions
Normalizing flows define a probability distribution by an explicit invertible transformation $\boldsymbol{\mathbf{z}}=f(\boldsymbol{\mathbf{x}})$. In this work, we present implicit normalizing flows (ImpFlows), which generalize normalizing flows by allowing the mapping to be implicitly defined by the roots of an equation $F(\boldsymbol{\mathbf{z}}, \boldsymbol{\mathbf{x}})= \boldsymbol{\mathbf{0}}$. ImpFlows build on residual flows (ResFlows) with a proper balance between expressiveness and tractability. Through theoretical analysis, we show that the function space of ImpFlow is strictly richer than that of ResFlows. Furthermore, for any ResFlow with a fixed number of blocks, there exists some function that ResFlow has a non-negligible approximation error. However, the function is exactly representable by a single-block ImpFlow. We propose a scalable algorithm to train and draw samples from ImpFlows. Empirically, we evaluate ImpFlow on several classification and density modeling tasks, and ImpFlow outperforms ResFlow with a comparable amount of parameters on all the benchmarks.
|
false |
DQSGD: DYNAMIC QUANTIZED STOCHASTIC GRADIENT DESCENT FOR COMMUNICATION-EFFICIENT DISTRIBUTED LEARNING
Distributed Learning Communication Gradient Quantization
Gradient quantization is widely adopted to mitigate communication costs in distributed learning systems. Existing gradient quantization algorithms often rely on design heuristics and/or empirical evidence to tune the quantization strategy for different learning problems. To the best of our knowledge, there is no theoretical framework characterizing the trade-off between communication cost and model accuracy under dynamic gradient quantization strategies. This paper addresses this issue by proposing a novel dynamic quantized SGD (DQSGD) framework, which enables us to optimize the quantization strategy for each gradient descent step by exploring the trade-off between communication cost and modeling error. In particular, we derive an upper bound, tight in some cases, of the modeling error for arbitrary dynamic quantization strategy. By minimizing this upper bound, we obtain an enhanced quantization algorithm with significantly improved modeling error under given communication overhead constraints. Besides, we show that our quantization scheme achieves a strengthened communication cost and model accuracy trade-off in a wide range of optimization models. Finally, through extensive experiments on large-scale computer vision and natural language processing tasks on CIFAR-10, CIFAR-100, and AG-News datasets, respectively. we demonstrate that our quantization scheme significantly outperforms the state-of-the-art gradient quantization methods in terms of communication costs.
|
false |
Active Tuning
Signal Filtering Recurrent Neural Network Time Series Denoising Temporal Gradients
We introduce Active Tuning, a novel paradigm for optimizing the internal dynamics of recurrent neural networks (RNNs) on the fly. In contrast to the conventional sequence-to-sequence mapping scheme, Active Tuning decouples the RNN's recurrent neural activities from the input stream, using the unfolding temporal gradient signal to tune the internal dynamics into the data stream. As a consequence, the model output depends only on its internal hidden dynamics and the closed-loop feedback of its own predictions; its hidden state is continuously adapted by means of the temporal gradient resulting from backpropagating the discrepancy between the signal observations and the model outputs through time. In this way, Active Tuning infers the signal actively but indirectly based on the originally learned temporal patterns, fitting the most plausible hidden state sequence into the observations. We demonstrate the effectiveness of Active Tuning on several time series prediction benchmarks, including multiple super-imposed sine waves, a chaotic double pendulum, and spatiotemporal wave dynamics. Active Tuning consistently improves the robustness, accuracy, and generalization abilities of all evaluated models. Moreover, networks trained for signal prediction and denoising can be successfully applied to a much larger range of noise conditions with the help of Active Tuning. Thus, given a capable time series predictor, Active Tuning enhances its online signal filtering, denoising, and reconstruction abilities without the need for additional training.
|
true |
Minigo: A Case Study in Reproducing Reinforcement Learning Research
case study reinforcement minigo research research minigo reproducibility key challenge area field results groundbreaking algorithm
The reproducibility of reinforcement-learning research has been highlighted as a key challenge area in the field. In this paper, we present a case study in reproducing the results of one groundbreaking algorithm, AlphaZero, a reinforcement learning system that learns how to play Go at a superhuman level given only the rules of the game. We describe Minigo, a reproduction of the AlphaZero system using publicly available Google Cloud Platform infrastructure and Google Cloud TPUs. The Minigo system includes both the central reinforcement learning loop as well as auxiliary monitoring and evaluation infrastructure. With ten days of training from scratch on 800 Cloud TPUs, Minigo can play evenly against LeelaZero and ELF OpenGo, two of the strongest publicly available Go AIs. We discuss the difficulties of scaling a reinforcement learning system and the monitoring systems required to understand the complex interplay of hyperparameter configurations.
|
false |
Veracity: An Online, Open-Source Fact-Checking Solution
Misinformation Fact-Checking AI for good Trust
The proliferation of misinformation poses a significant threat to society, exacerbated by the capabilities of generative AI.
This demo paper introduces Veracity, an open-source AI system designed to empower individuals to combat misinformation through transparent and accessible fact-checking. Veracity leverages the synergy between Large Language Models (LLMs) and web retrieval agents to analyze user-submitted claims and provide grounded veracity assessments with intuitive explanations. Key features include multilingual support, numerical scoring of claim veracity, and an interactive interface inspired by familiar messaging applications. This paper will showcase Veracity's ability to not only detect misinformation but also explain its reasoning, fostering media literacy and promoting a more informed society.
|
true |
Language Models Use Trigonometry to Do Addition
LLMs Mechanistic Interpretability Mathematics Reasoning
Mathematical reasoning is an increasingly important indicator of large language model (LLM) capabilities, yet we lack understanding of how LLMs process even simple mathematical tasks. To address this, we reverse engineer how three mid-sized LLMs compute addition. We first discover that numbers are represented in these LLMs as a generalized helix, which is strongly causally implicated for the tasks of addition and subtraction, and is also causally relevant for integer division, multiplication, and modular arithmetic. We then propose that LLMs compute addition by manipulating this generalized helix using the “Clock” algorithm: to solve $a+b$, the helices for $a$ and $b$ are manipulated to produce the $a+b$ answer helix which is then read out to model logits. We model influential MLP outputs, attention head outputs, and even individual neuron preactivations with these helices and verify our understanding with causal interventions. By demonstrating that LLMs represent numbers on a helix and manipulate this helix to perform addition, we present the first representation-level explanation of an LLM's mathematical capability.
|
true |
Generalization bounds via distillation
Generalization statistical learning theory theory distillation
This paper theoretically investigates the following empirical phenomenon: given a high-complexity network with poor generalization bounds, one can distill it into a network with nearly identical predictions but low complexity and vastly smaller generalization bounds. The main contribution is an analysis showing that the original network inherits this good generalization bound from its distillation, assuming the use of well-behaved data augmentation. This bound is presented both in an abstract and in a concrete form, the latter complemented by a reduction technique to handle modern computation graphs featuring convolutional layers, fully-connected layers, and skip connections, to name a few. To round out the story, a (looser) classical uniform convergence analysis of compression is also presented, as well as a variety of experiments on cifar and mnist demonstrating similar generalization performance between the original network and its distillation.
|
true |
Convex Regularization behind Neural Reconstruction
neural networks image reconstruction denoising interpretability robustness neural reconstruction convex duality inverse problems sparsity convex optimization
Neural networks have shown tremendous potential for reconstructing high-resolution images in inverse problems. The non-convex and opaque nature of neural networks, however, hinders their utility in sensitive applications such as medical imaging. To cope with this challenge, this paper advocates a convex duality framework that makes a two-layer fully-convolutional ReLU denoising network amenable to convex optimization. The convex dual network not only offers the optimum training with convex solvers, but also facilitates interpreting training and prediction. In particular, it implies training neural networks with weight decay regularization induces path sparsity while the prediction is piecewise linear filtering. A range of experiments with MNIST and fastMRI datasets confirm the efficacy of the dual network optimization problem.
|
false |
Unconditional Synthesis of Complex Scenes Using a Semantic Bottleneck
Unconditional Image Synthesis Complex Scene GAN Semantic Bottleneck
Coupling the high-fidelity generation capabilities of label-conditional image synthesis methods with the flexibility of unconditional generative models, we propose a semantic bottleneck GAN model for unconditional synthesis of complex scenes. We assume pixel-wise segmentation labels are available during training and use them to learn the scene structure through an unconditional progressive segmentation generation network. During inference, our model first synthesizes a realistic segmentation layout from scratch, then synthesizes a realistic scene conditioned on that layout through a conditional segmentation-to-image synthesis network. When trained end-to-end, the resulting model outperforms state-of-the-art generative models in unsupervised image synthesis on two challenging domains in terms of the Frechet Inception Distance and perceptual evaluations. Moreover, we demonstrate that the end-to-end training significantly improves the segmentation-to-image synthesis sub-network, which results in superior performance over the state-of-the-art when conditioning on real segmentation layouts.
|
false |
Robust Imitation via Decision-Time Planning
imitation learning reinforcement learning inverse reinforcement learning
The goal of imitation learning is to mimic expert behavior from demonstrations, without access to an explicit reward signal. A popular class of approaches infers the (unknown) reward function via inverse reinforcement learning (IRL) followed by maximizing this reward function via reinforcement learning (RL). The policies learned via these approaches are however very brittle in practice and deteriorate quickly even with small test-time perturbations due to compounding errors. We propose Imitation with Planning at Test-time (IMPLANT), a new algorithm for imitation learning that utilizes decision-time planning to correct for compounding errors of any base imitation policy. In contrast to existing approaches, we retain both the imitation policy and the rewards model at decision-time, thereby benefiting from the learning signal of the two components. Empirically, we demonstrate that IMPLANT significantly outperforms benchmark imitation learning approaches on standard control environments and excels at zero-shot generalization when subject to challenging perturbations in test-time dynamics.
|
false |
Understanding Self-supervised Learning with Dual Deep Networks
self-supervised learning teacher-student setting theoretical analysis hierarchical models representation learning
We propose a novel theoretical framework to understand self-supervised learning methods that employ dual pairs of deep ReLU networks (e.g., SimCLR, BYOL). First, we prove that in each SGD update of SimCLR, the weights at each layer are updated by a \emph{covariance operator} that specifically amplifies initial random selectivities that vary across data samples but survive averages over data augmentations. We show this leads to the emergence of hierarchical features, if the input data are generated from a hierarchical latent tree model. With the same framework, we also show analytically that in BYOL, the combination of BatchNorm and a predictor network creates an implicit contrastive term, acting as an approximate covariance operator. Additionally, for linear architectures we derive exact solutions for BYOL that provide conceptual insights into how BYOL can learn useful non-collapsed representations without any contrastive terms that separate negative pairs. Extensive ablation studies justify our theoretical findings.
|
false |
Regularization Cocktails for Tabular Datasets
deep learning regularization hyperparameter optimization benchmarks.
The regularization of prediction models is arguably the most crucial ingredient that allows Machine Learning solutions to generalize well on unseen data. Several types of regularization are popular in the Deep Learning community (e.g., weight decay, drop-out, early stopping, etc.), but so far these are selected on an ad-hoc basis, and there is no systematic study as to how different regularizers should be combined into the best “cocktail”. In this paper, we fill this gap, by considering the cocktails of 13 different regularization methods and framing the question of how to best combine them as a standard hyperparameter optimization problem. We perform a large-scale empirical study on 40 tabular datasets, concluding that, firstly, regularization cocktails substantially outperform individual regularization methods, even if the hyperparameters of the latter are carefully tuned; secondly, the optimal regularization cocktail depends on the dataset; and thirdly, regularization cocktails yield the state-of-the-art in classifying tabular datasets by outperforming Gradient-Boosted Decision Trees.
|
false |
Learning To Avoid Negative Transfer in Few Shot Transfer Learning
few shot learning negative transfer cubic spline ensemble learning
Many tasks in natural language understanding require learning relationships between two sequences for various tasks such as natural language inference, paraphrasing and entailment. These aforementioned tasks are similar in nature, yet they are often modeled individually. Knowledge transfer can be effective for closely related tasks, which is usually carried out using parameter transfer in neural networks. However, transferring all parameters, some of which irrelevant for a target task, can lead to sub-optimal results and can have a negative effect on performance, referred to as \textit{negative} transfer.
Hence, this paper focuses on the transferability of both instances and parameters across natural language understanding tasks by proposing an ensemble-based transfer learning method in the context of few-shot learning.
Our main contribution is a method for mitigating negative transfer across tasks when using neural networks, which involves dynamically bagging small recurrent neural networks trained on different subsets of the source task/s. We present a straightforward yet novel approach for incorporating these networks to a target task for few-shot learning by using a decaying parameter chosen according to the slope changes of a smoothed spline error curve at sub-intervals during training.
Our proposed method show improvements over hard and soft parameter sharing transfer methods in the few-shot learning case and shows competitive performance against models that are trained given full supervision on the target task, from only few examples.
|
true |
Interactive Image Generation Using Scene Graphs
Generative Models Image Generation Adversarial Learning Scene Graphs Interactive Graph Convolutional Network Image Translation Cascade Refinement Network
Recent years have witnessed some exciting developments in the domain of generating images from scene-based text descriptions. These approaches have primarily focused on generating images from a static text description and are limited to generating images in a single pass. They are unable to generate an image interactively based on an incrementally additive text description (something that is more intuitive and similar to the way we describe an image).
We propose a method to generate an image incrementally based on a sequence of graphs of scene descriptions (scene-graphs). We propose a recurrent network architecture that preserves the image content generated in previous steps and modifies the cumulative image as per the newly provided scene information. Our model utilizes Graph Convolutional Networks (GCN) to cater to variable-sized scene graphs along with Generative Adversarial image translation networks to generate realistic multi-object images without needing any intermediate supervision during training. We experiment with Coco-Stuff dataset which has multi-object images along with annotations describing the visual scene and show that our model significantly outperforms other approaches on the same dataset in generating visually consistent images for incrementally growing scene graphs.
|
true |
Collective Robustness Certificates: Exploiting Interdependence in Graph Neural Networks
Robustness certificates Adversarial robustness Graph neural networks
In tasks like node classification, image segmentation, and named-entity recognition we have a classifier that simultaneously outputs multiple predictions (a vector of labels) based on a single input, i.e. a single graph, image, or document respectively. Existing adversarial robustness certificates consider each prediction independently and are thus overly pessimistic for such tasks. They implicitly assume that an adversary can use different perturbed inputs to attack different predictions, ignoring the fact that we have a single shared input. We propose the first collective robustness certificate which computes the number of predictions that are simultaneously guaranteed to remain stable under perturbation, i.e. cannot be attacked. We focus on Graph Neural Networks and leverage their locality property - perturbations only affect the predictions in a close neighborhood - to fuse multiple single-node certificates into a drastically stronger collective certificate. For example, on the Citeseer dataset our collective certificate for node classification increases the average number of certifiable feature perturbations from $7$ to $351$.
|
false |
Active Learning in CNNs via Expected Improvement Maximization
active learning batch-mode active learning deep learning convolutional neural networks supervised learning regression classification MC dropout computer vision computational biology
Deep learning models such as Convolutional Neural Networks (CNNs) have demonstrated high levels of effectiveness in a variety of domains, including computer vision and more recently, computational biology. However, training effective models often requires assembling and/or labeling large datasets, which may be prohibitively time-consuming or costly. Pool-based active learning techniques have the potential to mitigate these issues, leveraging models trained on limited data to selectively query unlabeled data points from a pool in an attempt to expedite the learning process. Here we present "Dropout-based Expected IMprOvementS" (DEIMOS), a flexible and computationally-efficient approach to active learning that queries points that are expected to maximize the model's improvement across a representative sample of points. The proposed framework enables us to maintain a prediction covariance matrix capturing model uncertainty, and to dynamically update this matrix in order to generate diverse batches of points in the batch-mode setting. Our active learning results demonstrate that DEIMOS outperforms several existing baselines across multiple regression and classification tasks taken from computer vision and genomics.
|
true |
Generalized Variational Continual Learning
vcl gvcl variational continual learning baselines variational continual continual learning deals training models new tasks datasets online fashion
Continual learning deals with training models on new tasks and datasets in an online fashion. One strand of research has used probabilistic regularization for continual learning, with two of the main approaches in this vein being Online Elastic Weight Consolidation (Online EWC) and Variational Continual Learning (VCL). VCL employs variational inference, which in other settings has been improved empirically by applying likelihood-tempering. We show that applying this modification to VCL recovers Online EWC as a limiting case, allowing for interpolation between the two approaches. We term the general algorithm Generalized VCL (GVCL). In order to mitigate the observed overpruning effect of VI, we take inspiration from a common multi-task architecture, neural networks with task-specific FiLM layers, and find that this addition leads to significant performance gains, specifically for variational methods. In the small-data regime, GVCL strongly outperforms existing baselines. In larger datasets, GVCL with FiLM layers outperforms or is competitive with existing baselines in terms of accuracy, whilst also providing significantly better calibration.
|
true |
Efficient GPU-Accelerated Global Optimization for Inverse Problems
Scientific Machine Learning Inverse Problems Global Optimization GPU Computing
This paper introduces a novel hybrid multi-start optimization strategy for solving inverse problems involving nonlinear dynamical systems and machine learning architectures, accelerated by GPU computing on both NVIDIA and AMD GPUs. The method combines Particle Swarm Optimization (PSO) and the Limited-memory Broyden–Fletcher–Goldfarb–Shanno (L-BFGS) algorithms to address the challenges in parameter estimation for nonlinear dynamical systems. This hybrid strategy aims to leverage the global search capability of PSO and the efficient local convergence of L-BFGS. We experimentally show faster convergence by a factor of up to $8-30\times$ in a few non-convex problems with loss landscapes characterized by multiple local minima, which can cause regular optimization approaches to fail.
|
true |
Optimism in Reinforcement Learning with Generalized Linear Function Approximation
reinforcement learning optimism exploration function approximation theory regret analysis provable sample efficiency
We design a new provably efficient algorithm for episodic reinforcement learning with generalized linear function approximation. We analyze the algorithm under a new expressivity assumption that we call ``optimistic closure,'' which is strictly weaker than assumptions from prior analyses for the linear setting. With optimistic closure, we prove that our algorithm enjoys a regret bound of $\widetilde{O}\left(H\sqrt{d^3 T}\right)$ where $H$ is the horizon, $d$ is the dimensionality of the state-action features and $T$ is the number of episodes. This is the first statistically and computationally efficient algorithm for reinforcement learning with generalized linear functions.
|
false |
On the Robustness of Sentiment Analysis for Stock Price Forecasting
adversarial machine learning adversarial examples stock price forecasting finance
Machine learning (ML) models are known to be vulnerable to attacks both at training and test time. Despite the extensive literature on adversarial ML, prior efforts focus primarily on applications of computer vision to object recognition or sentiment analysis to movie reviews. In these settings, the incentives for adversaries to manipulate the model's prediction are often unclear and attacks require extensive control of direct inputs to the model. This makes it difficult to evaluate how severe the impact of vulnerabilities exposed is on systems deploying ML with little provenance guarantees for the input data. In this paper, we study adversarial ML with stock price forecasting. Adversarial incentives are clear and may be quantified experimentally through a simulated portfolio. We replicate an industry standard pipeline, which performs a sentiment analysis of Twitter data to forecast trends in stock prices. We show that an adversary can exploit the lack of provenance to indirectly use tweets to manipulate the model's perceived sentiment about a target company and in turn force the model to forecast price erroneously. Our attack is mounted at test time and does not modify the training data. Given past market anomalies, we conclude with a series of recommendations for the use of machine learning as input signal to trading algorithms.
|
true |
Temporally Sparse Attack for Fooling Large Language Models in Time Series Forecasting
Large Language model time series forecasting adversarial attack
Large Language Models (LLMs) have shown great potential in time series forecasting by capturing complex temporal patterns. Recent research reveals that LLM-based forecasters are highly sensitive to small input perturbations. However, existing attack methods often require modifying the entire time series, which is impractical in real-world scenarios. To address this, we propose a Temporally Sparse Attack (TSA) for LLM-based time series forecasting. By modeling the attack process as a Cardinality-Constrained Optimization Problem (CCOP), we develop a Subspace Pursuit (SP)--based method that restricts perturbations to a limited number of time steps, enabling efficient attacks. Experiments on advanced LLM-based time series models, including LLMTime (GPT-3.5, GPT-4, LLaMa, and Mistral), TimeGPT, and TimeLLM, show that modifying just 10\% of the input can significantly degrade forecasting performance across diverse datasets. This finding reveals a critical vulnerability in current LLM-based forecasters to low-dimensional adversarial attacks. Furthermore, our study underscores the practical application of CCOP and SP techniques in trustworthy AI, demonstrating their effectiveness in generating sparse, high-impact attacks and providing valuable insights into improving the robustness of AI systems.
|
true |
Syntax-Directed Variational Autoencoder for Structured Data
generative model for structured data syntax-directed generation molecule and program optimization variational autoencoder
Deep generative models have been enjoying success in modeling continuous data. However it remains challenging to capture the representations for discrete structures with formal grammars and semantics, e.g., computer programs and molecular structures. How to generate both syntactically and semantically correct data still remains largely an open problem. Inspired by the theory of compiler where syntax and semantics check is done via syntax-directed translation (SDT), we propose a novel syntax-directed variational autoencoder (SD-VAE) by introducing stochastic lazy attributes. This approach converts the offline SDT check into on-the-fly generated guidance for constraining the decoder. Comparing to the state-of-the-art methods, our approach enforces constraints on the output space so that the output will be not only syntactically valid, but also semantically reasonable. We evaluate the proposed model with applications in programming language and molecules, including reconstruction and program/molecule optimization. The results demonstrate the effectiveness in incorporating syntactic and semantic constraints in discrete generative models, which is significantly better than current state-of-the-art approaches.
|
false |
On Relating "Why?" and "Why Not?" Explanations
Explanability contrastive explanations duality
Explanations of Machine Learning (ML) models often address a ‘Why?’ question. Such explanations can be related with selecting feature-value pairs which are sufficient for the prediction. Recent work has investigated explanations that address
a ‘Why Not?’ question, i.e. finding a change of feature values that guarantee a change of prediction. Given their goals, these two forms of explaining predictions of ML models appear to be mostly unrelated. However, this paper demonstrates otherwise, and establishes a rigorous formal relationship between ‘Why?’ and ‘Why Not?’ explanations. Concretely, the paper proves that, for any given instance, ‘Why?’ explanations are minimal hitting sets of ‘Why Not?’ explanations and vice-versa. Furthermore, the paper devises novel algorithms for extracting and enumerating both forms of explanations.
|
false |
Demystifying Learning of Unsupervised Neural Machine Translation
Unsupervised Neural Machine Translation Marginal Likelihood Maximization Mutual Information
Unsupervised Neural Machine Translation or UNMT has received great attention
in recent years. Though tremendous empirical improvements have been achieved,
there still lacks theory-oriented investigation and thus some fundamental
questions like \textit{why} certain training protocol can work or not under
\textit{what} circumstances have not yet been well understood. This paper
attempts to provide theoretical insights for the above questions. Specifically,
following the methodology of comparative study, we leverage two perspectives,
i) \textit{marginal likelihood maximization} and ii) \textit{mutual information}
from information theory, to understand the different learning effects from the
standard training protocol and its variants. Our detailed analyses reveal
several critical conditions for the successful training of UNMT.
|
false |
Learning the Step-size Policy for the Limited-Memory Broyden-Fletcher-Goldfarb-Shanno Algorithm
Unconstrained optimization Step-size policy L-BFGS Learned optimizers
We consider the problem of how to learn a step-size policy for the Limited-Memory Broyden-Fletcher-Goldfarb-Shanno (L-BFGS) algorithm. This is a limited computational memory quasi-Newton method widely used for deterministic unconstrained optimization but currently avoided in large-scale problems for requiring step sizes to be provided at each iteration. Existing methodologies for the step size selection for L-BFGS use heuristic tuning of design parameters and massive re-evaluations of the objective function and gradient to find appropriate step-lengths. We propose a neural network architecture with local information of the current iterate as the input. The step-length policy is learned from data of similar optimization problems, avoids additional evaluations of the objective function, and guarantees that the output step remains inside a pre-defined interval. The corresponding training procedure is formulated as a stochastic optimization problem using the backpropagation through time algorithm. The performance of the proposed method is evaluated on the training of classifiers for the MNIST database for handwritten digits and for CIFAR-10. The results show that the proposed algorithm outperforms heuristically tuned optimizers such as ADAM, RMSprop, L-BFGS with a backtracking line search and L-BFGS with a constant step size. The numerical results also show that a learned policy can be used as a warm-start to train new policies for different problems after a few additional training steps, highlighting its potential use in multiple large-scale optimization problems.
|
true |
MDCROW: AUTOMATING MOLECULAR DYNAMICS WORKFLOWS WITH LARGE LANGUAGE MODELS
agent agentic AI computational biology molecular dynamics large language models
Molecular dynamics (MD) simulations are essential for understanding biomolecular systems but remain challenging to automate. Recent advances in large language models (LLM) have demonstrated success in automating complex scientific tasks using LLM-based agents. In this paper, we introduce MDCrow, an agentic LLM assistant capable of automating MD workflows. MDCrow uses chain-of-thought reasoning over 40 expert-designed tools for handling and processing files, setting up simulations, analyzing the simulation outputs, and retrieving relevant information from literature and databases. We assess MDCrow's performance across 25 tasks of varying complexity, and we evaluate the agent's robustness to both task complexity and prompt style. GPT-4o is able to complete complex tasks with low variance, followed closely by Llama3-405b, a compelling open-source model. While prompt style does not influence the best models' performance, it may improve performance on smaller models.
|
true |
CleanGen: Mitigating Backdoor Attacks for Generation Tasks in Large Language Models
Inference methods interactive and collaborative generation
The remarkable performance of large language models (LLMs) in generation tasks has enabled practitioners to leverage publicly available models to power custom applications, such as chatbots and virtual assistants. However, the data used to train or fine-tune these LLMs is often undisclosed, allowing an attacker to compromise the data and inject backdoors into the models. In this paper, we develop a novel inference time defense, named CleanGen, to mitigate backdoor attacks for generation tasks in LLMs. CleanGen is a lightweight and effective decoding strategy that is compatible with the state-of-the-art (SOTA) LLMs. Our insight behind CleanGen is that compared to other LLMs, backdoored LLMs assign significantly higher probabilities to tokens representing the attacker-desired contents. These discrepancies in token probabilities enable CleanGen to identify suspicious tokens favored by the attacker and replace them with tokens generated by another LLM that is not compromised by the same attacker, thereby avoiding generation of attacker-desired content. We evaluate CleanGen against five SOTA backdoor attacks. Our results show that CleanGen achieves lower attack success rates (ASR) compared to five SOTA baseline defenses for all five backdoor attacks. Moreover, LLMs deploying CleanGen maintain helpfulness in their responses when serving benign user queries with minimal added computational overhead.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.