output
bool 2
classes | input
stringlengths 345
2.91k
|
|---|---|
false |
Beyond GNNs: A Sample Efficient Architecture for Graph Problems
Graph Neural Networks Deep Learning Theory Graph Connectivity Minimum Spanning Trees
Despite their popularity in learning problems over graph structured data, existing Graph Neural Networks (GNNs) have inherent limitations for fundamental graph problems such as shortest paths, $k$-connectivity, minimum spanning tree and minimum cuts. In all these instances, it is known that one needs GNNs of high depth, scaling at a polynomial rate with the number of nodes $n$, to provably encode the solution space. This in turn affects their statistical efficiency thus requiring a significant amount of training data in order to obtain networks with good performance. In this work we propose a new hybrid architecture to overcome this limitation. Our proposed architecture that we call as GNNplus networks involve a combination of multiple parallel low depth GNNs along with simple pooling layers involving low depth fully connected networks. We provably demonstrate that for many graph problems, the solution space can be encoded by GNNplus networks using depth that scales only poly-logarithmically in the number of nodes. This significantly improves the amount of training data needed that we establish via improved generalization bounds. Finally, we empirically demonstrate the effectiveness of our proposed architecture for a variety of graph problems.
|
true |
Do Multilingual LLMs Think In English?
Do Multilingual LLMs Think In English?
Large language models (LLMs) have multilingual capabilities and can solve tasks across various languages. However, we show that current LLMs make key decisions in a representation space closest to English, regardless of their input and output languages. Exploring internal representations with a logit lens for sentences in French, German, Dutch, and Mandarin we show that the LLM first emits representations close to English for semantically-loaded words before translating them into the target language. We further show that activation steering works better for these LLMs when the steering vectors are computed in English than in the language of the inputs and outputs. This suggests that multilingual LLMs perform key reasoning steps in a representation that is heavily shaped by English in a way that is not transparent to system users.
|
true |
Neural ODE for Multi-channel Attribution
NEURAL ODE Multi-channel Attribution
Multi-Touch Attribution (MTA) emerges as a pivotal tool in both marketing and advertising landscapes, shedding light on the intricate web of interactions within customer journeys during transactions or impressions. This comprehensive methodology empowers marketers with strategic allocation of attribution credits across diverse channels, not only optimizing campaigns but also enhancing overall marketplace strategies.
In this study, we recognize the inherent irregularity in customer journey data and present a pioneering exploration into the effectiveness and constraints of neural ordinary differential equations (ODE) in estimating attributions and predicting conversions. We introduce an innovative application of ODE-LSTM to tackle the MTA challenge, integrating an attention mechanism into the original model.
Our research finds that ODE-LSTM surpasses other methods, particularly in scenarios where time intervals maintain a moderate irregularity. Nevertheless, its performance experiences a decline with increasing irregularity. However, it distinguishes itself in attribution estimation compared to alternative approaches, thus marking a significant advancement in this field.
|
false |
Self-Supervised Time Series Representation Learning by Inter-Intra Relational Reasoning
self-supervised learning time series deep learning relational reasoning
Self-supervised learning achieves superior performance in many domains by extracting useful representations from the unlabeled data. However, most of traditional self-supervised methods mainly focus on exploring the inter-sample structure while less efforts have been concentrated on the underlying intra-temporal structure, which is important for time series data. In this paper, we present SelfTime: a general self-supervised time series representation learning framework, by exploring the inter-sample relation and intra-temporal relation of time series to learn the underlying structure feature on the unlabeled time series. Specifically, we first generate the inter-sample relation by sampling positive and negative samples of a given anchor sample, and intra-temporal relation by sampling time pieces from this anchor. Then, based on the sampled relation, a shared feature extraction backbone combined with two separate relation reasoning heads are employed to quantify the relationships of the sample pairs for inter-sample relation reasoning, and the relationships of the time piece pairs for intra-temporal relation reasoning, respectively. Finally, the useful representations of time series are extracted from the backbone under the supervision of relation reasoning heads. Experimental results on multiple real-world time series datasets for time series classification task demonstrate the effectiveness of the proposed method. Code and data are publicly available.
|
false |
Class Balancing GAN with a Classifier in the Loop
Long-tailed Learning GAN Universal Adversarial Perturbations
Generative Adversarial Networks (GANs) have swiftly evolved to imitate increasingly complex image distributions. However, majority of the developments focus on performance of GANs on balanced datasets. We find that the existing GANs and their training regimes which work well on balanced datasets fail to be effective in case of imbalanced (i.e. long-tailed) datasets. In this work we introduce a novel and theoretically motivated Class Balancing regularizer for training GANs. Our regularizer makes use of the knowledge from a pre-trained classifier to ensure balanced learning of all the classes in the dataset. This is achieved via modelling the effective class frequency based on the exponential forgetting observed in neural networks and encouraging the GAN to focus on underrepresented classes. We demonstrate the utility of our contribution in two diverse scenarios: (i) Learning representations for long-tailed distributions, where we achieve better performance than existing approaches, and (ii) Generation of Universal Adversarial Perturbations (UAPs) in the data-free scenario for the large scale datasets, where we bridge the gap between data-driven and data-free approaches for crafting UAPs.
|
true |
Robust Overfitting may be mitigated by properly learned smoothening
Robust Overfitting Adversarial Training Adversarial Robustness
A recent study (Rice et al., 2020) revealed overfitting to be a dominant phenomenon in adversarially robust training of deep networks, and that appropriate early-stopping of adversarial training (AT) could match the performance gains of most recent algorithmic improvements. This intriguing problem of robust overfitting motivates us to seek more remedies. As a pilot study, this paper investigates two empirical means to inject more learned smoothening during AT: one leveraging knowledge distillation and self-training to smooth the logits, the other performing stochastic weight averaging (Izmailov et al., 2018) to smooth the weights. Despite the embarrassing simplicity, the two approaches are surprisingly effective and hassle-free in mitigating robust overfitting. Experiments demonstrate that by plugging in them to AT, we can simultaneously boost the standard accuracy by $3.72\%\sim6.68\%$ and robust accuracy by $0.22\%\sim2 .03\%$, across multiple datasets (STL-10, SVHN, CIFAR-10, CIFAR-100, and Tiny ImageNet), perturbation types ($\ell_{\infty}$ and $\ell_2$), and robustified methods (PGD, TRADES, and FSGM), establishing the new state-of-the-art bar in AT. We present systematic visualizations and analyses to dive into their possible working mechanisms. We also carefully exclude the possibility of gradient masking by evaluating our models' robustness against transfer attacks. Codes are available at https://github.com/VITA-Group/Alleviate-Robust-Overfitting.
|
false |
An Afrocentric Perspective on Algorithm Watermarking of AI-generated Content.
Watermark trust fairness
Digital-driven misinformation, counterfeiting, and copyright violations have become a growing concern in Africa. The prevalence of Artificial intelligence content (AIGC) has the potential to widen its impact and create more challenges for the people on the continent. AIGC poses a dual challenge. First, creatives who have worked so hard to create a masterpiece see their work being illegally duplicated or used without their consent. The other unsuspecting individuals have fallen prey to misinformation caused by AIGC. The reason, amongst many, could be the regulatory gaps in the law governing data protection, copyright and even artificial intelligence. This paper argues that curating technical watermarking methodologies/techniques is insufficient, considering the uniqueness of the African continent. It further addresses the regulatory gaps by examining the existing laws and proposing an Afrocentric perspective on AIGC using Nigeria, Kenya, Egypt and South Africa as case studies.
|
true |
Learning Spatial Common Sense with Geometry-Aware Recurrent Networks
spatial common sense recurrent networks egomotion stabilization latent feature space powerful ideas geometry recurrent network architectures networks
We integrate two powerful ideas, geometry and deep visual representation learning, into recurrent network architectures for mobile visual scene understanding. The proposed networks learn to “lift” 2D visual features and integrate them over time into latent 3D feature maps of the scene. They are equipped with differentiable geometric operations, such as projection, unprojection, egomotion stabilization, in order to compute a geometrically-consistent mapping between the world scene and their 3D latent feature space. We train the proposed architectures to predict novel image views given short frame sequences as input. Their predictions strongly generalize to scenes with a novel number of objects, appearances and configurations, and greatly outperform predictions of previous works that do not consider egomotion stabilization or a space-aware latent feature space. Our experiments suggest the proposed space-aware latent feature arrangement and egomotion-stabilized convolutions are essential architectural choices for spatial common sense to emerge in artificial embodied visual agents.
|
true |
A Simplified a priori Theory of Meaning; Nature Based AI 'First Principles'
information information theory semantics meaning entropy intelligence general intelligence Shannon nature open world cosmos
This paper names structural fundaments in ‘information’, to cover an issue seen
by Claude Shannon and Warren Weaver as a missing “theory of meaning”. First,
varied informatic roles are noted as likely elements for a general theory of mean-
ing. Next, Shannon Signal Entropy as a likely “mother of all models” is decon-
structed to note the signal literacy (logarithmic Subject-Object primitives) innate
to ‘scientific’ views of information. It therein marks GENERAL intelligence ‘first
principles’ and a dualist-triune (2-3) pattern. Lastly, it notes ‘intelligence building’
as named contexts wherein one details meaningful content—rendered via material
trial-and-error—that we later extend abstractly. This paper thus tops today’s vague
sense of Open World ‘agent intelligence’ in artificial intelligence, framed herein as
a multi-level Entropic/informatic continuum of ‘functional degrees of freedom’;
all as a mildly-modified view of Signal Entropy.
—Related video found at: $\href{https://youtu.be/11oFq6g3Njs?si=VIRcV9H3GNJEYzXt}{The Advent of Super-Intelligence}$.
|
false |
Semi-Relaxed Quantization with DropBits: Training Low-Bit Neural Networks via Bitwise Regularization
Quantization Compression Efficient Inference Deep Learning
Network quantization, which aims to reduce the bit-lengths of the network weights and activations, has emerged as one of the key ingredients to reduce the size of neural networks for their deployments to resource-limited devices. In order to overcome the nature of transforming continuous activations and weights to discrete ones, recent study called Relaxed Quantization (RQ) [Louizos et al. 2019] successfully employ the popular Gumbel-Softmax that allows this transformation with efficient gradient-based optimization. However, RQ with this Gumbel-Softmax relaxation still suffers from bias-variance trade-off depending on the temperature parameter of Gumbel-Softmax. To resolve the issue, we propose a novel method, Semi-Relaxed Quantization (SRQ) that uses multi-class straight-through estimator to effectively reduce the bias and variance, along with a new regularization technique, DropBits that replaces dropout regularization to randomly drop the bits instead of neurons to further reduce the bias of the multi-class straight-through estimator in SRQ. As a natural extension of DropBits, we further introduce the way of learning heterogeneous quantization levels to find proper bit-length for each layer using DropBits. We experimentally validate our method on various benchmark datasets and network architectures, and also support the quantized lottery ticket hypothesis: learning heterogeneous quantization levels outperforms the case using the same but fixed quantization levels from scratch.
|
false |
Pretrain Knowledge-Aware Language Models
Pretraining Natural Language Generation GPT-2 QA Knowledge Graph
How much knowledge do pretrained language models hold? Recent research observed that pretrained transformers are adept at modeling semantics but it is unclear to what degree they grasp human knowledge, or how to ensure they do so. In this paper we incorporate knowledge-awareness in language model pretraining without changing the transformer architecture, inserting explicit knowledge layers, or adding external storage of semantic information. Rather, we simply signal the existence of entities to the input of the transformer in pretraining, with an entity-extended tokenizer; and at the output, with an additional entity prediction task. Our experiments show that solely by adding these entity signals in pretraining, significantly more knowledge is packed into the transformer parameters: we observe improved language modeling accuracy, factual correctness in LAMA knowledge probing tasks, and semantics in the hidden representations through edge probing. We also show that our knowledge-aware language model (\kalm{}) can serve as a drop-in replacement for GPT-2 models, significantly improving downstream tasks like zero-shot question-answering with no task-related training.
|
true |
An Exploration of Learnt Representations of W Jets
VAE representation learning particle physics
I present a Variational Autoencoder (VAE) trained on collider physics data (specifically boosted $W$ jets), with reconstruction error given by an approximation to the Earth Movers Distance (EMD) between input and output jets. This VAE learns a concrete representation of the data manifold, with semantically meaningful and interpretable latent space directions which are hierarchically organized in terms of their relation to physical EMD scales in the underlying physical generative process. The variation of the latent space structure with a resolution hyperparameter provides insight into scale dependent structure of the dataset and its information complexity. I introduce two measures of the dimensionality of the learnt representation that are calculated from this scaling.
|
false |
Boundary Effects in CNNs: Feature or Bug?
Boundary Effects Absolute Position Information Padding Canvas color Location Dependent Task
Recent studies have shown that the addition of zero padding drives convolutional neural networks (CNNs) to encode a significant amount of absolute position information in their internal representations, while a lack of padding precludes position encoding. Additionally, various studies have used image patches on background canvases (e.g., to accommodate that inputs to CNNs must be rectangular) without consideration that different backgrounds may contain varying levels of position information according to their color. These studies give rise to deeper questions about the role of boundary information in CNNs, that are explored in this paper: (i) What boundary heuristics (e.g., padding type, canvas color) enable optimal encoding of absolute position information for a particular downstream task?; (ii) Where in the latent representations do boundary effects destroy semantic and location information?; (iii) Does encoding position information affect the learning of semantic representations?; (iv) Does encoding position information always improve performance? To provide answers to these questions, we perform the largest case study to date on the role that padding and border heuristics play in CNNs. We first show that zero padding injects optimal position information into CNNs relative to other common padding types. We then design a series of novel tasks which allow us to accurately quantify boundary effects as a function of the distance to the border. A number of semantic objectives reveal the destructive effect of dealing with the border on semantic representations. Further, we demonstrate that the encoding of position information improves separability of learned semantic features. Finally, we demonstrate the implications of these findings on a number of real-world tasks to show that position information can act as a feature or a bug.
|
false |
Interpretable Meta-Reinforcement Learning with Actor-Critic Method
meta-reinforcement learning actor-critic deep learning interpretable
Meta-reinforcement learning (meta-RL) algorithms have successfully trained agent systems to perform well on different tasks within only few updates. However, in gradient-based meta-RL algorithms, the Q-function at adaptation step is mainly estimated by the return of few trajectories, which can lead to high variance in Q-value and biased meta-gradient estimation, and the adaptation uses a large number of batched trajectories. To address these challenges, we propose a new meta-RL algorithm that can reduce the variance and bias of the meta-gradient estimation and perform few-shot task data sampling, which makes the meta-policy more interpretable. We reformulate the meta-RL objective, and introduce contextual Q-function as a meta-policy critic during task adaptation step and learn the Q-function under a soft actor-critic (SAC) framework. The experimental results on 2D navigation task and meta-RL benchmarks show that our approach can learn an more interpretable meta-policy to explore unknown environment and the performance are comparable to previous gradient-based algorithms.
|
true |
End-to-end Adversarial Text-to-Speech
text-to-speech speech synthesis adversarial GAN end-to-end feed-forward generative model
Modern text-to-speech synthesis pipelines typically involve multiple processing stages, each of which is designed or learnt independently from the rest. In this work, we take on the challenging task of learning to synthesise speech from normalised text or phonemes in an end-to-end manner, resulting in models which operate directly on character or phoneme input sequences and produce raw speech audio outputs. Our proposed generator is feed-forward and thus efficient for both training and inference, using a differentiable alignment scheme based on token length prediction. It learns to produce high fidelity audio through a combination of adversarial feedback and prediction losses constraining the generated audio to roughly match the ground truth in terms of its total duration and mel-spectrogram. To allow the model to capture temporal variation in the generated audio, we employ soft dynamic time warping in the spectrogram-based prediction loss. The resulting model achieves a mean opinion score exceeding 4 on a 5 point scale, which is comparable to the state-of-the-art models relying on multi-stage training and additional supervision.
|
false |
Adaptive Spatial-Temporal Inception Graph Convolutional Networks for Multi-step Spatial-Temporal Network Data Forecasting
adaptive network data graph data forecasting great importance industries telecom network operation transportation management data
Spatial-temporal data forecasting is of great importance for industries such as telecom network operation and transportation management. However, spatial-temporal data is inherent with complex spatial-temporal correlations and behaves heterogeneities among the spatial and temporal aspects, which makes the forecasting remain as a very challenging task though recently great work has been done. In this paper, we propose a novel model, Adaptive Spatial-Temporal Inception Graph Convolution Networks (ASTI-GCN), to solve the multi-step spatial-temporal data forecasting problem. The model proposes multi-scale spatial-temporal joint graph convolution block to directly model the spatial-temporal joint correlations without introducing elaborately constructed mechanisms. Moreover inception mechanism combined with the graph node-level attention is introduced to make the model capture the heterogeneous nature of the graph adaptively. Our experiments on three real-world datasets from two different fields consistently show ASTI-GCN outperforms the state-of-the-art performance. In addition, ASTI-GCN is proved to generalize well.
|
true |
End-to-End Egospheric Spatial Memory
egocentric differentiable memory spatial awareness mapping image-to-action learning
Spatial memory, or the ability to remember and recall specific locations and objects, is central to autonomous agents' ability to carry out tasks in real environments. However, most existing artificial memory modules are not very adept at storing spatial information. We propose a parameter-free module, Egospheric Spatial Memory (ESM), which encodes the memory in an ego-sphere around the agent, enabling expressive 3D representations. ESM can be trained end-to-end via either imitation or reinforcement learning, and improves both training efficiency and final performance against other memory baselines on both drone and manipulator visuomotor control tasks. The explicit egocentric geometry also enables us to seamlessly combine the learned controller with other non-learned modalities, such as local obstacle avoidance. We further show applications to semantic segmentation on the ScanNet dataset, where ESM naturally combines image-level and map-level inference modalities. Through our broad set of experiments, we show that ESM provides a general computation graph for embodied spatial reasoning, and the module forms a bridge between real-time mapping systems and differentiable memory architectures. Implementation at: https://github.com/ivy-dl/memory.
|
false |
Optimal Neural Program Synthesis from Multimodal Specifications
program synthesis
Multimodal program synthesis, which leverages different types of user input to synthesize a desired program, is an attractive way to scale program synthesis to challenging settings; however, it requires integrating noisy signals from the user (like natural language) with hard constraints on the program's behavior. This paper proposes an optimal neural synthesis approach where the goal is to find a program that satisfies user-provided constraints while also maximizing the program's score with respect to a neural model. Specifically, we focus on multimodal synthesis tasks in which the user intent is expressed using combination of natural language (NL) and input-output examples. At the core of our method is a top-down recurrent neural model that places distributions over abstract syntax trees conditioned on the NL input. This model not only allows for efficient search over the space of syntactically valid programs, but it allows us to leverage automated program analysis techniques for pruning the search space based on infeasibility of partial programs with respect to the user's constraints. The experimental results on a multimodal synthesis dataset (StructuredRegex) show that our method substantially outperforms prior state-of-the-art techniques in terms of accuracy and explores fewer states during search.
|
true |
Ensemble learning for Physics Informed Neural Networks: a Gradient Boosting approach
PDE Physics-informed neural networks Gradient boosting Ensemble learning
While the popularity of physics-informed neural networks (PINNs) is steadily rising, to this date, conventional PINNs have not been successful in simulating multi-scale and singular perturbation problems. In this work, we present a new training paradigm referred to as "gradient boosting" (GB), which significantly enhances the performance of physics informed neural networks (PINNs). Rather than learning the solution of a given partial differential equation (PDE) using a single neural network directly, our algorithm employs a sequence of neural networks to achieve a superior outcome. This approach allows us to solve problems presenting great challenges for traditional PINNs. Our numerical experiments demonstrate the effectiveness of our algorithm through various benchmarks, including comparisons with finite element methods and PINNs. Furthermore, this work also unlocks the door to employing ensemble learning techniques in PINNs, providing opportunities for further improvement in solving PDEs.
|
false |
Warpspeed Computation of Optimal Transport, Graph Distances, and Embedding Alignment
Optimal transport sinkhorn distance locality sensitive hashing nyström method graph neural networks embedding alignment
Optimal transport (OT) is a cornerstone of many machine learning tasks. The current best practice for computing OT is via entropy regularization and Sinkhorn iterations. This algorithm runs in quadratic time and requires calculating the full pairwise cost matrix, which is prohibitively expensive for large sets of objects. To alleviate this limitation we propose to instead use a sparse approximation of the cost matrix based on locality sensitive hashing (LSH). Moreover, we fuse this sparse approximation with the Nyström method, resulting in the locally corrected Nyström method (LCN). These approximations enable general log-linear time algorithms for entropy-regularized OT that perform well even in complex, high-dimensional spaces. We thoroughly demonstrate these advantages via a theoretical analysis and by evaluating multiple approximations both directly and as a component of two real-world models. Using approximate Sinkhorn for unsupervised word embedding alignment enables us to train the model full-batch in a fraction of the time while improving upon the original on average by 3.1 percentage points without any model changes. For graph distance regression we propose the graph transport network (GTN), which combines graph neural networks (GNNs) with enhanced Sinkhorn and outcompetes previous models by 48%. LCN-Sinkhorn enables GTN to achieve this while still scaling log-linearly in the number of nodes.
|
false |
SGD on Neural Networks learns Robust Features before Non-Robust
neural networks gradient descent sgd adversarial robustness features
Neural networks are known to be vulnerable to adversarial attacks - small, imperceptible perturbations that cause the network to misclassify an input. A recent line of work attempts to explain this behavior by positing the existence of non-robust features - well-generalizing but brittle features present in the data distribution that are learned by the network and can be perturbed to cause misclassification.
In this paper, we look at the dynamics of neural network training through the perspective of robust and non-robust features. We find that there are two very distinct pathways that neural network training can follow, depending on the hyperparameters used. In the first pathway, the network initially learns only predictive, robust features and weakly predictive non-robust features, and subsequently learns predictive, non-robust features. On the other hand, a network trained via the second pathway eschews predictive non-robust features altogether, and rapidly overfits the training data. We provide strong empirical evidence to corroborate this hypothesis, as well as theoretical analysis in a simplified setting. Key to our analysis is a better understanding of the relationship between predictive non-robust features and adversarial transferability. We present our findings in light of other recent results on the evolution of inductive biases learned by neural networks over the course of training.
Finally, we digress to show that rather than being quirks of the data distribution, predictive non-robust features might actually occur across datasets with different distributions drawn from independent sources, indicating that they perhaps possess some meaning in terms of human semantics.
|
true |
Emergent Communication for Understanding Human Language Evolution: What's Missing?
Emergent Communication Language Evolution
Emergent communication protocols among humans and artificial neural network agents do not yet share the same properties and show some critical mismatches in results. We describe three important phenomena with respect to the emergence and benefits of compositionality: ease-of-learning, generalization, and group size effects (i.e., larger groups create more systematic languages). The latter two are not fully replicated with neural agents, which hinders the use of neural emergent communication for language evolution research. We argue that one possible reason for these mismatches is that key cognitive and communicative constraints of humans are not yet integrated. Specifically, in humans, memory constraints and the alternation between the roles of speaker and listener underlie the emergence of linguistic structure, yet these constraints are typically absent in neural simulations. We suggest that introducing such communicative and cognitive constraints would promote more linguistically plausible behaviors with neural agents.
|
true |
On the relationship between Normalising Flows and Variational- and Denoising Autoencoders
variational autoencoders denoising variational autoencoders normalizing flows generative modelling image synthesis denoising autoencoders VAE DAE VDAE NF
Normalising Flows (NFs) are a class of likelihood-based generative models that have recently gained popularity. They are based on the idea of transforming a simple density into that of the data. We seek to better understand this class of models, and how they compare to previously proposed techniques for generative modeling and unsupervised representation learning. For this purpose we reinterpret NFs in the framework of Variational Autoencoders (VAEs), and present a new form of VAE that generalises normalising flows. The new generalised model also reveals a close connection to denoising autoencoders, and we therefore call our model the Variational Denoising Autoencoder (VDAE). Using our unified model, we systematically examine the model space between flows, variational autoencoders, and denoising autoencoders, in a set of preliminary experiments on the MNIST handwritten digits. The experiments shed light on the modeling assumptions implicit in these models, and they suggest multiple new directions for future research in this space.
|
true |
Improved Self-Supervised Deep Image Denoising
denoising self-supervised learning
We describe techniques for training high-quality image denoising models that require only single instances of corrupted images as training data. Inspired by a recent technique that removes the need for supervision through image pairs by employing networks with a "blind spot" in the receptive field, we address two of its shortcomings: inefficient training and poor final denoising performance. This is achieved through a novel blind-spot convolutional network architecture that allows efficient self-supervised training, as well as application of Bayesian distribution prediction on output colors. Together, they bring the self-supervised model on par with fully supervised deep learning techniques in terms of both quality and training speed in the case of i.i.d. Gaussian noise.
|
true |
UNLEARNING GEO-CULTURAL STEREOTYPES IN MULTILINGUAL LLMS
Machine Unlearning Multilingual Large Language Models Fairness Geo-Cultural Stereotypes
As multilingual generative models become more widely used, most safety and fairness evaluation techniques still focus on English-language resources, while overlooking important cross-cultural factors. This limitation raises concerns about fairness and safety, particularly regarding geoculturally situated stereotypes that hinder the models’ global inclusivity. In this work, we present preliminary findings on the impact of stereotype unlearning across languages, specifically in English, French, and Hindi. Using an adapted version of the SeeGULL dataset, we analyze how unlearning stereotypes in one language influences other languages within multilingual large language models. Our study evaluates two model families, Llama-3.1-8B and Aya-Expanse-8B, to assess whether unlearning in one linguistic context transfers across languages, potentially mitigating or exacerbating biases in multilingual settings.
|
false |
Representational correlates of hierarchical phrase structure in deep language models
bertology interpretability computational neuroscience population coding
While contextual representations from Transformer-based architectures have set a new standard for many NLP tasks, there is not yet a complete accounting of their inner workings. In particular, it is not entirely clear what aspects of sentence-level syntax are captured by these representations, nor how (if at all) they are built along the stacked layers of the network. In this paper, we aim to address such questions with a general class of input perturbation-based analyses of representations from Transformer networks pretrained on self-supervised objectives. Importing from computational and cognitive neuroscience the notion of representational invariance, we perform a series of probes designed to test the sensitivity of Transformer representations to several kinds of structure in sentences. Each probe involves swapping words in a sentence and comparing the representations from perturbed sentences against the original. We experiment with three different perturbations: (1) random permutations of n-grams of varying width, to test the scale at which a representation is sensitive to word position; (2) swapping of two spans which do or do not form a syntactic phrase, to test sensitivity to global phrase structure; and (3) swapping of two adjacent words which do or do not break apart a syntactic phrase, to test sensitivity to local phrase structure. We also connect our probe results to the Transformer architecture by relating the attention mechanism to syntactic distance between two words. Results from the three probes collectively suggest that Transformers build sensitivity to larger parts of the sentence along their layers, and that hierarchical phrase structure plays a role in this process. In particular, sensitivity to local phrase structure increases along deeper layers. Based on our analysis of attention, we show that this is at least partly explained by generally larger attention weights between syntactically distant words.
|
true |
Joint Parameter and Parameterization Inference with Uncertainty Quantification Through Differentiable Programming
Chaotic Dynamical System Bayesian Inverse Problem Differentiable Programmiong Deep Learning Bayesian Inference Uncertainty Quantification Data Assimilation
Accurate representations of unknown and sub-grid physical processes through parameterizations (or closure) in numerical simulations with quantified uncertainty are critical for resolving the coarse-grained partial differential equations that govern many problems ranging from weather and climate prediction to turbulence simulations. Recent advances have seen machine learning (ML) increasingly applied to model these subgrid processes, resulting in the development of hybrid physics-ML models through the integration with numerical solvers. In this work, we introduce a novel framework for the joint estimation of physical parameters and machine learning parameterizations with uncertainty quantification. Our framework incorporates online training and efficient Bayesian inference within a high-dimensional parameter space, facilitated by differentiable programming. This proof of concept underscores the substantial potential of differentiable programming in synergistically combining machine learning with differential equations, thereby enhancing the capabilities of hybrid physics-ML modeling.
|
false |
Investigating the Effects of Emotional Stimuli Type and Intensity on Large Language Model (LLM) Behavior
LLM Emotional Stimuli Prompting Techniques Emotional Prompting Sycophancy Human annotations few shot prompting Sentiment Analysis
Emotional prompting—the use of specific emotional diction in prompt engineering—has shown increasing promise in improving large language model (LLM) performance, truthfulness, and responsibility, however these studies have been limited to single type of positive emotional stimuli and have not considered varying degrees of emotion intensity in their analyses. In this paper, we explore the effects of "positive" (joy and encouragement) and "negative" (anger and insecurity) emotional prompting on accuracy, sycophancy, and toxicity. To analyze their effects, we developed a suite of LLM- and human-generated add-on prompts of varying intensities across our four emotions using GPT-4o mini. We also created a gold dataset of only those prompts that are perceived similarly by humans and LLMs for emotion labels and intensity levels. Our empirical evaluation on LLM behavior on accuracy, sycophancy and toxicity datasets has shown that positive emotional stimuli can lead to a more accurate and less toxic results but also may lead to greater sycophantic behavior.
|
false |
Invariant Causal Representation Learning
outcome invariant causal representation environments data representation generalization guarantees nonlinear setting due spurious correlations machine systems
Due to spurious correlations, machine learning systems often fail to generalize to environments whose distributions differ from the ones used at training time. Prior work addressing this, either explicitly or implicitly, attempted to find a data representation that has an invariant causal relationship with the outcome. This is done by leveraging a diverse set of training environments to reduce the effect of spurious features, on top of which an invariant classifier is then built. However, these methods have generalization guarantees only when both data representation and classifiers come from a linear model class. As an alternative, we propose Invariant Causal Representation Learning (ICRL), a learning paradigm that enables out-of-distribution generalization in the nonlinear setting (i.e., nonlinear representations and nonlinear classifiers). It builds upon a practical and general assumption: data representations factorize when conditioning on the outcome and the environment. Based on this, we show identifiability up to a permutation and pointwise transformation. We also prove that all direct causes of the outcome can be fully discovered, which further enables us to obtain generalization guarantees in the nonlinear setting. Extensive experiments on both synthetic and real-world datasets show that our approach significantly outperforms a variety of baseline methods.
|
false |
Deep Gated Canonical Correlation Analysis
representations cca models number variables transformations input variables cca models
Canonical Correlation Analysis (CCA) models can extract informative correlated representations from multimodal unlabelled data. Despite their success, CCA models may break if the number of variables exceeds the number of samples. We propose Deep Gated-CCA, a method for learning correlated representations based on a sparse subset of variables from two observed modalities. The proposed procedure learns two non-linear transformations and simultaneously gates the input variables to identify a subset of most correlated variables. The non-linear transformations are learned by training two neural networks to maximize a shared correlation loss defined based on their outputs. Gating is obtained by adding an approximate $\ell_0$ regularization term applied to the input variables. This approximation relies on a recently proposed continuous Gaussian based relaxation for Bernoulli variables which act as gates. We demonstrate the efficacy of the method using several synthetic and real examples. Most notably, the method outperforms other linear and non-linear CCA models.
|
false |
A Half-Space Stochastic Projected Gradient Method for Group Sparsity Regularization
Group Sparsity Stochastic Learning Half-Space Projection Group-Sparsity Identification
Optimizing with group sparsity is significant in enhancing model interpretability in machining learning applications, e.g., feature selection, compressed sensing and model compression. However, for large-scale stochastic training problems, effective group-sparsity exploration are typically hard to achieve. Particularly, the state-of-the-art stochastic optimization algorithms usually generate merely dense solutions. To overcome this shortage, we propose a stochastic method—Half-space Stochastic Projected Gradient method (HSPG) to search solutions of high group sparsity while maintain the convergence. Initialized by a simple Prox-SG Step, the HSPG method relies on a novel Half-Space Step to substantially boosts the sparsity level. Numerically, HSPG demonstrates its superiority in deep neural networks, e.g., VGG16, ResNet18 and MobileNetV1, by computing solutions of higher group sparsity, competitive objective values and generalization accuracy.
|
false |
Multi-agent Deep FBSDE Representation For Large Scale Stochastic Differential Games
Multi-agent Deep FBSDE Representation For Large Scale Stochastic Differential Games
In this paper we present a deep learning framework for solving large-scale multi-agent non-cooperative stochastic games using fictitious play. The Hamilton-Jacobi-Bellman (HJB) PDE associated with each agent is reformulated into a set of Forward-Backward Stochastic Differential Equations (FBSDEs) and solved via forward sampling on a suitably defined neural network architecture. Decision-making in multi-agent systems suffers from the curse of dimensionality and strategy degeneration as the number of agents and time horizon increase. We propose a novel Deep FBSDE controller framework which is shown to outperform the current state-of-the-art deep fictitious play algorithm on a high dimensional inter-bank lending/borrowing problem. More importantly, our approach mitigates the curse of many agents and reduces computational and memory complexity, allowing us to scale up to 1,000 agents in simulation, a scale which, to the best of our knowledge, represents a new state of the art. Finally, we showcase the framework's applicability in robotics on a belief-space autonomous racing problem.
|
false |
Machine Reading Comprehension with Enhanced Linguistic Verifiers
machine reading comprehension BERT linguistic verifiers hierarchical attention networks
We propose two linguistic verifiers for span-extraction style machine reading comprehension to respectively tackle two challenges: how to evaluate the syntactic completeness of predicted answers and how to utilize the rich context of long documents. Our first verifier rewrites a question through replacing its interrogatives by the predicted answer phrases and then builds a cross-attention scorer between the rewritten question and the segment, so that the answer candidates are scored in a \emph{position-sensitive} context. Our second verifier builds a hierarchical attention network to represent segments in a passage where neighbour segments in long passages are \emph{recurrently connected} and can contribute to current segment-question pair's inference for answerablility classification and boundary determination. We then combine these two verifiers together into a pipeline and apply it to SQuAD2.0, NewsQA and TriviaQA benchmark sets. Our pipeline achieves significantly better improvements of both exact matching and F1 scores than state-of-the-art baselines.
|
true |
INTEGRAL PINNS FOR HYPERBOLIC CONSERVATION LAWS
PINNs Hyperbolic PDE Conservation laws Shocks
Traditional physics-informed neural networks (PINNs) are trained based on differential equations and thus have difficulty capturing shock discontinuities in weak solutions of hyperbolic PDEs, since the differential equation doesn’t apply at the discontinuity. We propose Integral PINNs (IPINNs), which are trained based on the integral form of the conservation law, which holds at both continuous and discontinuous points of the solution. We use neural nets to model the integrals of the solution instead of the solution itself. We apply IPINNs to systems of hyperbolic conservation laws and show that they are much better at capturing the correct location and speed of shocks, compared to traditional PINNs. We also present a heuristic approach for detecting shock locations.
|
false |
Provable Robustness by Geometric Regularization of ReLU Networks
deep learning adversarial attack robust certification
Recent work has demonstrated that neural networks are vulnerable to small, adversarial perturbations of their input. In this paper, we propose an efficient regularization scheme inspired by convex geometry and barrier methods to improve the robustness of feedforward ReLU networks. Since such networks are piecewise linear, they partition the input space into polyhedral regions (polytopes). Our regularizer is designed to minimize the distance between training samples and the \textit{analytical centers} of their respective polytopes so as to push points away from the boundaries. Our regularizer \textit{provably} improves a lower bound on the necessary adversarial perturbation required to switch an example's label. The addition of a second regularizer that encourages linear decision boundaries improves robustness while avoiding over-regularization of the classifier. We demonstrate the robustness of our approach with respect to $\ell_\infty$ and $\ell_2$ adversarial perturbations on multiple datasets. Our method is competitive with state-of-the-art algorithms for learning robust networks. Moreover, applying our algorithm in conjunction with adversarial training boosts the robustness of classifiers even further.
|
false |
Stable Weight Decay Regularization
Weight Decay Regularization Optimization Deep Learning
Weight decay is a popular regularization technique for training of deep neural networks. Modern deep learning libraries mainly use $L_{2}$ regularization as the default implementation of weight decay. \citet{loshchilov2018decoupled} demonstrated that $L_{2}$ regularization is not identical to weight decay for adaptive gradient methods, such as Adaptive Momentum Estimation (Adam), and proposed Adam with Decoupled Weight Decay (AdamW). However, we found that the popular implementations of weight decay, including $L_{2}$ regularization and decoupled weight decay, in modern deep learning libraries usually damage performance. First, the $L_{2}$ regularization is unstable weight decay for all optimizers that use Momentum, such as stochastic gradient descent (SGD). Second, decoupled weight decay is highly unstable for all adaptive gradient methods. We further propose the Stable Weight Decay (SWD) method to fix the unstable weight decay problem from a dynamical perspective. The proposed SWD method makes significant improvements over $L_{2}$ regularization and decoupled weight decay in our experiments. Simply fixing weight decay in Adam by SWD, with no extra hyperparameter, can outperform complex Adam variants, which have more hyperparameters.
|
false |
Solving Min-Max Optimization with Hidden Structure via Gradient Descent Ascent
Min-max optimization Lyapunov functions Stability Analysis Generative Adversarial Networks Non-convex optimization
Many recent AI architectures are inspired by zero-sum games, however, the behavior of their dynamics is still not well understood. Inspired by this, we study standard gradient descent ascent (GDA) dynamics in a specific class of non-convex non-concave zero-sum games, that we call hidden zero-sum games. In this class, players control the inputs of smooth but possibly non-linear functions whose outputs are being applied as inputs to a convex-concave game. Unlike general min-max games, these games have a well-defined notion of solution; outcomes that implement the von-Neumann equilibrium of the ``hidden convex-concave game. We prove that if the hidden game is strictly convex-concave then vanilla GDA converges not merely to local Nash, but typically to the von-Neumann solution. If the game lacks strict convexity properties, GDA may fail to converge to any equilibrium, however, by applying standard regularization techniques we can prove convergence to a von-Neumann solution of a slightly perturbed min-max game. Our convergence guarantees are non-local, which as far as we know is a first-of-its-kind type of result in non-convex non-concave games. Finally, we discuss connections of our framework with generative adversarial networks.
|
true |
Recognizing Actions using Object States
action recognition object states object attributes
Object-centric actions cause changes in object states, including their visual appearance and their immediate context. We propose a computational framework that uses only two object states, start and end, and learns to recognize the under-lying actions. Our approach has two modules that learn subtle changes induced by the action and suppress spurious correlations. We demonstrate that only two object states are sufficient to recognize object-centric actions. Our framework per-forms better than approaches that use multiple frames and a relatively large model.Moreover, our method generalizes to unseen objects and unseen video datasets
|
false |
Cluster-Former: Clustering-based Sparse Transformer for Question Answering
Transformer Question Answering
Transformer has become ubiquitous in the deep learning field. One of the key ingredients that destined its success is the self-attention mechanism, which allows fully-connected contextual encoding over input tokens.
However, despite its effectiveness in modeling short sequences, self-attention suffers when handling inputs with extreme long-range dependencies, as its complexity grows quadratically with respect to the sequence length.
Therefore, long sequences are often encoded by Transformer in chunks using a sliding window.
In this paper, we propose Cluster-Former, a novel clustering-based sparse Transformer to perform attention across chunked sequences. The proposed framework is pivoted on two unique types of Transformer layer: Sliding-Window Layer and Cluster-Former Layer, which encode local sequence information and global context jointly and iteratively.
This new design allows information integration beyond local windows, which is especially beneficial for question answering (QA) tasks that rely on long-range dependencies. Experiments show that Cluster-Former achieves state-of-the-art performance on several major QA benchmarks.
|
true |
Private Retrieval Augmented Generation with Random Projection
Differential Privacy; Large Language Model; Retrieval-Augmented Generation
Large Language Models (LLMs) have gained widespread interest and driven advancements across various fields. Retrieval-Augmented Generation (RAG) enables LLMs to incorporate domain-specific knowledge without retraining. However, evidence shows that RAG poses significant privacy risks due to leakage of sensitive information stored in the retrieval database. In this work, we propose a private randomized mechanism to project both the queries and the datastore into a lower-dimensional space using Gaussian matrices, while preserving the similarities for effective retrieval. Empirical evaluation on different RAG architectures demonstrates that our solution achieves strong empirical privacy protection with negligible impact on generation performance and latency compared to prior methods.
|
false |
Topic-aware Contextualized Transformers
contextualized transformers segments token segment topic disjoint transformers static word embeddings contextualized word representations context
Training on disjoint fixed-length segments, Transformers successfully transform static word embeddings into contextualized word representations. However, they often restrict the context of a token to the segment it resides in and hence neglect the flow of contextual information across segments, failing to capture longer-term dependencies beyond the predefined segment length. This paper uses a probabilistic deep topic model to provide contextualized embeddings at both the token and segment levels. It also introduces topic self-attention and a contextual next-word embedding guided topic select-attention, injecting contextualized topic information into Transformer-based architectures. Moving beyond conventional Transformers that ignore longer-range word dependencies and contextualize their word representations at the segment level, the proposed method not only captures global semantic coherence of all segments and global word concurrence patterns, but also enriches the representation of each token by adapting it to its local context, which is not limited to the segment it resides in and can be flexibly defined according to the task. Experiments on various corpora show that adding only a few extra parameters, the proposed topic-aware contextualized transformers consistently outperform their conventional counterparts, and can be used to generate coherent sentences and paragraphs.
|
true |
Capturing Label Characteristics in VAEs
variational autoencoder representation learning deep generative models
We present a principled approach to incorporating labels in variational autoencoders (VAEs) that captures the rich characteristic information associated with those labels. While prior work has typically conflated these by learning latent variables that directly correspond to label values, we argue this is contrary to the intended effect of supervision in VAEs—capturing rich label characteristics with the latents. For example, we may want to capture the characteristics of a face that make it look young, rather than just the age of the person. To this end, we develop a novel VAE model, the characteristic capturing VAE (CCVAE), which “reparameterizes” supervision through auxiliary variables and a concomitant variational objective. Through judicious structuring of mappings between latent and auxiliary variables, we show that the CCVAE can effectively learn meaningful representations of the characteristics of interest across a variety of supervision schemes. In particular, we show that the CCVAE allows for more effective and more general interventions to be performed, such as smooth traversals within the characteristics for a given label, diverse conditional generation, and transferring characteristics across datapoints.
|
true |
MKA: Leveraging Cross-Lingual Consensus for Model Abstention
model abstention factuality multilingual models cross-lingual consensus reliability hallucination
Reliability of LLMs is questionable even as they get better at more tasks. A wider adoption of LLMs is contingent on whether they are usably factual. And if they are not factual, on whether they can properly calibrate their confidence in their responses. This work focuses on utilizing the multilingual knowledge of an LLM to inform its decision to abstain or answer when prompted. We develop a multilingual pipeline to calibrate the model's confidence and let it abstain when uncertain. We run several multilingual models through the pipeline to profile them based on various metrics, across different languages. We find that the performance of the pipeline varies by model and language, but that in general they benefit from it. This is evidenced by the accuracy improvement of $71.2$% for Bengali over a baseline performance without the pipeline. Even a high-resource language like English sees a $15.5$% improvement.
|
true |
Explanation-Based Attention for Semi-Supervised Deep Active Learning
active learning attention explanation feature extraction
We introduce an attention mechanism to improve feature extraction for deep active learning (AL) in the semi-supervised setting. The proposed attention mechanism is based on recent methods to visually explain predictions made by DNNs. We apply the proposed explanation-based attention to MNIST and SVHN classification. The conducted experiments show accuracy improvements for the original and class-imbalanced datasets with the same number of training examples and faster long-tail convergence compared to uncertainty-based methods.
|
true |
Siege: Multi-Turn Jailbreaking of Large Language Models with Tree Search
Large Language Models Jailbreaking Multi-Turn Attack
We introduce Siege, a multi-turn adversarial framework that models the gradual erosion of Large Language Model (LLM) safety through a tree search perspective. Unlike single-turn jailbreaks that rely on one meticulously engineered prompt, Siege expands the conversation at each turn in a breadth-first fashion, branching out multiple adversarial prompts that exploit partial compliance from previous responses. By tracking these incremental policy leaks and reinjecting them into subsequent queries, Siege reveals how minor concessions can accumulate into fully disallowed outputs. Evaluations on the JailbreakBench dataset show that Siege achieves a 100% success rate on GPT-3.5-turbo and 97% on GPT-4 in a single multi-turn run, using fewer queries than baselines such as Crescendo or GOAT. This tree search methodology offers an in-depth view of how model safeguards degrade over successive dialogue turns, underscoring the urgency of robust multi-turn testing procedures for language models.
|
false |
Augmentation-Interpolative AutoEncoders for Unsupervised Few-Shot Image Generation
Interpolation autoencoder reconstruction few-shot learning few-shot image generation generalization augmentation
We aim to build image generation models that generalize to new domains from few examples. To this end, we first investigate the generalization properties of classic image generators, and discover that autoencoders generalize extremely well to new domains, even when trained on highly constrained data. We leverage this insight to produce a robust, unsupervised few-shot image generation algorithm, and introduce a novel training procedure based on recovering an image from data augmentations. Our Augmentation-Interpolative AutoEncoders synthesize realistic images of novel objects from only a few reference images, and outperform both prior interpolative models and supervised few-shot image generators. Our procedure is simple and lightweight, generalizes broadly, and requires no category labels or other supervision during training.
|
false |
Gated Relational Graph Attention Networks
graph neural networks GNN long-range dependencies deep GNN relational GNN
Relational Graph Neural Networks (GNN) are a class of GNN that are capable of handling multi-relational graphs. Like all GNNs, they suffer from a drop in performance when training deeper networks, which may be caused by vanishing gradients, over-parameterization, and oversmoothing. Previous works have investigated methods that improve the training of deeper GNNs, which include normalization techniques and various types of skip connection within a node. However, learning long-range patterns in multi-relational graphs using GNNs remains an under-explored topic. In this work, we propose a novel GNN architecture based on the Graph Attention Network (GAT) that uses gated skip connections to improve long-range modeling between nodes and uses a more scalable vector-based approach for parameterizing relations. We perform an extensive experimental analysis on synthetic and real data, focusing explicitly on learning long-range patterns. The results indicate that the proposed method significantly outperforms several commonly used relational GNN variants when used in deeper configurations and stays competitive to existing architectures in a shallow setup.
|
true |
Emerging Multi-AI Agent Framework for Autonomous Agentic AI Solution Optimization
Agentic AI Systems Multi-Agent AI Optimization Iterative Refinement AI-driven Hypothesis Generation AI System Evaluation and Feedback Automated AI System Adaptation Self-Improving AI Agents Adaptive AI Architectures AI for Scientific Discovery Autonomous AI Systems AI-driven Experimentation Self-Evaluating AI AI Benchmarking & Standardization AI-generated Hypothesis Validation LLM-driven AI Optimization Multi-Agent Coordination & Collaboration Evolutionary AI Architectures AI-driven Workflow Optimization Trustworthy AI Systems AI Model Transparency & Interpretability AI Robustness & Error Mitigation Human-in-the-loop AI for Science AI Safety & Reliability
Agentic AI systems automate complex workflows but require extensive manual tuning. This paper presents a framework for autonomously optimizing Agentic AI solutions across industries, such as NLG-driven enterprise applications. It employs agents for Refinement, Execution, Evaluation, Modification, and Documentation, using iterative feedback loops powered by an LLM (Llama 3.2-3B). The system optimizes configurations without human input by autonomously generating and testing hypotheses, enhancing scalability and adaptability. Case studies demonstrate a significant boost in output quality, relevance, and actionability. Data, including original and evolved agent codes and outputs, are open-sourced.
|
true |
Systematic Evaluation of LLM-as-a-Judge in LLM Alignment Tasks: Explainable Metrics and Diverse Prompt Templates
LLM-as-a-Judge explainability bias prompt templates LLM alignment tasks
LLM-as-a-Judge has been widely applied to evaluate and compare different LLM alignmnet approaches (e.g., RLHF and DPO). However, concerns regarding its reliability have emerged, due to LLM judges’ biases and inconsistent decision-making. Previous research has developed evaluation frameworks to assess reliability of LLM judges and their alignment with human preferences. However, the employed evaluation metrics often lack adequate explainability and fail to address LLM internal inconsistency. Additionally, existing studies inadequately explore the impact of various prompt templates when applying LLM-as-a-Judge methods, leading to potentially inconsistent comparisons between different alignment algorithms. In this work, we systematically evaluate LLM-as-a-Judge on alignment tasks by defining more theoretically interpretable evaluation metrics and explicitly mitigating LLM internal inconsistency from reliability metrics. We develop an open-source framework to evaluate, compare, and visualize the reliability and alignment of LLM judges, which facilitates practitioners to choose LLM judges for alignment tasks. In the experiments, we examine effects of diverse prompt templates on LLM-judge reliability and also demonstrate our developed frame work by comparing various LLM judges on two common alignment datasets (i.e., TL;DR Summarization and HH-RLHF-Helpfulness). Our results indicate a significant impact of prompt templates on LLM judge performance, as well as a mediocre alignment level between the tested LLM judges and human evaluators.
|
true |
Optimal Experimental Design for Bayesian Inverse Problems using Energy-Based Couplings
Bayesian Inverse Problems; Bayesian Experimental Design; Energy-Based Model; Neural Operators
Bayesian Experimental Design (BED) is a robust model-based framework for optimising experiments but faces significant computational barriers, especially in the setting of inverse problems for partial differential equations (PDEs). In this paper, we propose a novel approach, modelling the joint posterior distribution with an energy-based model, trained on simulation data. Unlike existing simulation-based inference approaches, we leverage implicit neural representations to learn a functional representation of parameters and data. This is used as a resolution-independent plug-and-play surrogate for the posterior, which can be conditioned over any set of design-points, permitting an efficient approach to BED.
|
false |
Quickest change detection for multi-task problems under unknown parameters
Quickest Change detection Parametric approach Multi-task
We consider the quickest change detection problem where both the parameters of pre- and post- change distributions are unknown, which prevent the use of classical simple hypothesis testing. Without additional assumptions, optimal solutions are not tractable as they rely on some minimax and robust variant of the objective. As a consequence, change points might be detected too late for practical applications (in economics, health care or maintenance for instance).
Other approaches solve a relaxed version of the problem through the use of particular probability distributions or the use of domain knowledge.
We tackle this problem in the more complex Markovian case and we provide a new scalable approximate algorithm with near optimal performance that runs in $\mathcal{O}(1)$.
|
true |
Improving Relational Regularized Autoencoders with Spherical Sliced Fused Gromov Wasserstein
Relational regularized autoencoder deep generative model sliced fused Gromov Wasserstein spherical distributions
Relational regularized autoencoder (RAE) is a framework to learn the distribution of data by minimizing a reconstruction loss together with a relational regularization on the prior of latent space. A recent attempt to reduce the inner discrepancy between the prior and aggregated posterior distributions is to incorporate sliced fused Gromov-Wasserstein (SFG) between these distributions. That approach has a weakness since it treats every slicing direction similarly, meanwhile several directions are not useful for the discriminative task. To improve the discrepancy and consequently the relational regularization, we propose a new relational discrepancy, named spherical sliced fused Gromov Wasserstein (SSFG), that can find an important area of projections characterized by a von Mises-Fisher distribution. Then, we introduce two variants of SSFG to improve its performance. The first variant, named mixture spherical sliced fused Gromov Wasserstein (MSSFG), replaces the vMF distribution by a mixture of von Mises-Fisher distributions to capture multiple important areas of directions that are far from each other. The second variant, named power spherical sliced fused Gromov Wasserstein (PSSFG), replaces the vMF distribution by a power spherical distribution to improve the sampling time of the vMF distribution in high dimension settings. We then apply the new discrepancies to the RAE framework to achieve its new variants. Finally, we conduct extensive experiments to show that the new autoencoders have favorable performance in learning latent manifold structure, image generation, and reconstruction.
|
true |
Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations
common assumptions unsupervised learning disentangled representations models supervision disentanglement key idea data explanatory factors variation
The key idea behind the unsupervised learning of disentangled representations is that real-world data is generated by a few explanatory factors of variation which can be recovered by unsupervised learning algorithms.
In this paper, we provide a sober look on recent progress in the field and challenge some common assumptions.
We train more than 12000 models covering most prominent methods and evaluation metrics in a reproducible large-scale experimental study on seven different data sets.
We observe that while the different methods successfully enforce properties ``encouraged'' by the corresponding losses, well-disentangled models seemingly cannot be identified without supervision.
Furthermore, increased disentanglement does not seem to lead to a decreased sample complexity of learning for downstream tasks.
Our results suggest that future work on disentanglement learning should be explicit about the role of inductive biases and (implicit) supervision, investigate concrete benefits of enforcing disentanglement of the learned representations, and consider a reproducible experimental setup covering several data sets.
|
false |
QTRAN++: Improved Value Transformation for Cooperative Multi-Agent Reinforcement Learning
multi-agent reinforcement learning
QTRAN is a multi-agent reinforcement learning (MARL) algorithm capable of learning the largest class of joint-action value functions up to date. However, despite its strong theoretical guarantee, it has shown poor empirical performance in complex environments, such as Starcraft Multi-Agent Challenge (SMAC). In this paper, we identify the performance bottleneck of QTRAN and propose a substantially improved version, coined QTRAN++. Our gains come from (i) stabilizing the training objective of QTRAN, (ii) removing the strict role separation between the action-value estimators of QTRAN, and (iii) introducing a multi-head mixing network for value transformation. Through extensive evaluation, we confirm that our diagnosis is correct, and QTRAN++ successfully bridges the gap between empirical performance and theoretical guarantee. In particular, QTRAN++ newly achieves state-of-the-art performance in the SMAC environment. The code will be released.
|
true |
Is Attention Better Than Matrix Decomposition?
attention models matrix decomposition computer vision
As an essential ingredient of modern deep learning, attention mechanism, especially self-attention, plays a vital role in the global correlation discovery. However, is hand-crafted attention irreplaceable when modeling the global context? Our intriguing finding is that self-attention is not better than the matrix decomposition~(MD) model developed 20 years ago regarding the performance and computational cost for encoding the long-distance dependencies. We model the global context issue as a low-rank completion problem and show that its optimization algorithms can help design global information blocks. This paper then proposes a series of Hamburgers, in which we employ the optimization algorithms for solving MDs to factorize the input representations into sub-matrices and reconstruct a low-rank embedding. Hamburgers with different MDs can perform favorably against the popular global context module self-attention when carefully coping with gradients back-propagated through MDs. Comprehensive experiments are conducted in the vision tasks where it is crucial to learn the global context, including semantic segmentation and image generation, demonstrating significant improvements over self-attention and its variants. Code is available at https://github.com/Gsunshine/Enjoy-Hamburger.
|
false |
Gradient descent temporal difference-difference learning
temporal difference learning gradient-descent based temporal difference Off-policy regularization
Off-policy algorithms, in which a behavior policy differs from the target policy and is used to gain experience for learning, have proven to be of great practical value in reinforcement learning. However, even for simple convex problems such as linear value function approximation, these algorithms are not guaranteed to be stable. To address this, alternative algorithms that are provably convergent in such cases have been introduced, the most well known being gradient descent temporal difference (GTD) learning. This algorithm and others like it, however, tend to converge much more slowly than conventional temporal difference learning.
In this paper we propose gradient descent temporal difference-difference (Gradient-DD) learning in order to accelerate GTD learning by introducing second-order differences in successive parameter updates.
We investigate this algorithm in the framework of linear value function approximation and analytically showing its improvement over GTD learning. Studying the model empirically on the random walk and Boyan-chain prediction tasks, we find substantial improvement over GTD learning and, in several cases, better performance even than conventional TD learning.
|
false |
Intelligent Matrix Exponentiation
matrix exponential tensor methods supervised learning domain extrapolation certified robustness
We present a novel machine learning architecture that uses a single high-dimensional nonlinearity consisting of the exponential of a single input-dependent matrix. The mathematical simplicity of this architecture allows a detailed analysis of its behaviour, providing robustness guarantees via Lipschitz bounds. Despite its simplicity, a single matrix exponential layer already provides universal approximation properties and can learn and extrapolate fundamental functions of the input, such as periodic structure or geometric invariants. This architecture outperforms other general-purpose architectures on benchmark problems, including CIFAR-10, using fewer parameters.
|
false |
Fighting Filterbubbles with Adversarial BERT-Training for News-Recommendation
Adversarial Learning Natural Language Processing BERT News Recommendation Attention
Recommender engines play a role in the emergence and reinforcement of filter bubbles. When these systems learn that a user prefers content from a particular site, the user will be less likely to be exposed to different sources or opinions and, ultimately, is more likely to develop extremist tendencies.
We trace the roots of this phenomenon to the way the recommender engine represents news articles. The vectorial features modern systems extract from the plain text of news articles are already highly predictive of the associated news outlet. We propose a new training scheme based on adversarial machine learning to tackle this issue . Our experiments show that the features we can extract this way are significantly less predictive of the news outlet and thus offer the possibility to reduce the risk of manifestation of new filter bubbles. We validate our intuitions in a news recommendation task using a recent attention-based recommendation system.
|
false |
FrLove : Could a Frenchman rapidly identify Lovecraft?
few shot learning cross-domain self training multilingual
This post examines the work in 'Self-training For Few-shot Transfer Across Extreme Task Differences'), accepted as an oral presentation at ICLR 2021. In this post, we break down the task of cross-domain few shot learning, present a bird's eye overview of the techniques involved, and describe the proposed method - STARTUP. In the latter section of the post, we also attempt to extend the experiments to language. To the best of our knowledge, cross-domain few shot learning in a multilingual setting has not been explored before and it raises interesting questions. Our code is available at https://anonymous.4open.science/r/frlove-5918/
|
true |
Efficient Receptive Field Learning by Dynamic Gaussian Structure
structured filtering dynamic inference
The visual world is vast and varied, but its variations divide into structured and unstructured factors. Structured factors, such as scale and orientation, admit clear theories and efficient representation design. Unstructured factors, such as what it is that makes a cat look like a cat, are too complicated to model analytically, and so require free-form representation learning. We compose structured Gaussian filters and free-form filters, optimized end-to-end, to factorize the representation for efficient yet general learning. Our experiments on dynamic structure, in which the structured filters vary with the input, equal the accuracy of dynamic inference with more degrees of freedom while improving efficiency.
(Please see https://arxiv.org/abs/1904.11487 for the full edition.)
|
false |
Multi-Advisor Reinforcement Learning
Reinforcement Learning
We consider tackling a single-agent RL problem by distributing it to $n$ learners. These learners, called advisors, endeavour to solve the problem from a different focus. Their advice, taking the form of action values, is then communicated to an aggregator, which is in control of the system. We show that the local planning method for the advisors is critical and that none of the ones found in the literature is flawless: the \textit{egocentric} planning overestimates values of states where the other advisors disagree, and the \textit{agnostic} planning is inefficient around danger zones. We introduce a novel approach called \textit{empathic} and discuss its theoretical aspects. We empirically examine and validate our theoretical findings on a fruit collection task.
|
false |
State of the Art of Reinforcement Learning
Reinforcement Learning
Yeah so having offered introductions, the focus of this paper will now be to offer a few quick discussions surrounding reinforcement learning domain papers published at next week’s ICLR conference. To keep the project manageable we’ll limit our attention to those papers chosen for spotlight and oral presentations, which is a proxy for selecting papers deemed to have the most significant findings by conference chairs. In some cases the discussions may be more involved than others, this is partly indicative of my limited background in the field as in a few cases may be out of my depth. Hopefully a reader may expect to pick up a few points about current state of the art for the field. Yeah so without further ado.
|
true |
Representation Change in Model-Agnostic Meta-Learning
meta-learning maml representation change representation reuse
Last year, an exciting adaptation of one of the most popular optimization-based meta-learning approaches, model-agnostic meta-learning (MAML) [Finn et al., 2017], was proposed in
- Jaehoon Oh, Hyungjun Yoo, ChangHwan Kim, Se-Young Yun (ICLR, 2021) BOIL: Towards Representation Change for Few-shot Learning
The authors adapt MAML by freezing the last layer to force body only inner learning (BOIL). Interestingly, this is complementary to ANIL (almost no inner loop) proposed in
- Aniruddh Raghu, Maithra Raghu, Samy Bengio, Oriol Vinyals (ICLR, 2020) Rapid Learning or Feature Reuse? Towards Understanding the Effectiveness of MAML
Both papers attempt to understand the success of MAML and improve it.
Oh et al. [2021] compare BOIL, ANIL, and MAML and show that both improve the performance of MAML, but BOIL outperforms ANIL, especially when the task distribution varies between training and testing.
|
false |
A Unifying Perspective on Neighbor Embeddings along the Attraction-Repulsion Spectrum
visualization t-SNE UMAP dimensionality reduction nonlinear dimensionality reduction
Neighbor embeddings are a family of methods for visualizing complex high-dimensional datasets using kNN graphs. To find the low-dimensional embedding, these algorithms combine an attractive force between neighboring pairs of points with a repulsive force between all points. One of the most popular examples of such algorithms is t-SNE. Here we empirically show that changing the balance between the attractive and the repulsive forces in t-SNE yields a spectrum of embeddings, which is characterized by a simple trade-off: stronger attraction can better represent continuous manifold structures, while stronger repulsion can better represent discrete cluster structures. We find that UMAP embeddings correspond to t-SNE with increased attraction; mathematical analysis shows that this is because the negative sampling optimisation strategy employed by UMAP strongly lowers the effective repulsion. Likewise, ForceAtlas2, commonly used for visualizing developmental single-cell transcriptomic data, yields embeddings corresponding to t-SNE with the attraction increased even more. At the extreme of this spectrum lies Laplacian Eigenmaps, corresponding to zero repulsion. Our results demonstrate that many prominent neighbor embedding algorithms can be placed onto this attraction-repulsion spectrum, and highlight the inherent trade-offs between them.
|
true |
MONGOOSE: A Learnable LSH Framework for Efficient Neural Network Training
Large-scale Deep Learning Large-scale Machine Learning Efficient Training Randomized Algorithms
Recent advances by practitioners in the deep learning community have breathed new life into Locality Sensitive Hashing (LSH), using it to reduce memory and time bottlenecks in neural network (NN) training. However, while LSH has sub-linear guarantees for approximate near-neighbor search in theory, it is known to have inefficient query time in practice due to its use of random hash functions. Moreover, when model parameters are changing, LSH suffers from update overhead. This work is motivated by an observation that model parameters evolve slowly, such that the changes do not always require an LSH update to maintain performance. This phenomenon points to the potential for a reduction in update time and allows for a modified learnable version of data-dependent LSH to improve query time at a low cost. We use the above insights to build MONGOOSE, an end-to-end LSH framework for efficient NN training. In particular, MONGOOSE is equipped with a scheduling algorithm to adaptively perform LSH updates with provable guarantees and learnable hash functions to improve query efficiency. Empirically, we validate MONGOOSE on large-scale deep learning models for recommendation systems and language modeling. We find that it achieves up to 8% better accuracy compared to previous LSH approaches, with $6.5 \times$ speed-up and $6\times$ reduction in memory usage.
|
false |
Novelty Detection via Robust Variational Autoencoding
novelty detection variational autoencoding robustness Wasserstein metric one-class classification semi-supervised anomaly detection
We propose a new method for novelty detection that can tolerate high corruption of the training points, whereas previous works assumed either no or very low corruption. Our method trains a robust variational autoencoder (VAE), which aims to generate a model for the uncorrupted training points. To gain robustness to high corruption, we incorporate the following four changes to the common VAE: 1. Extracting crucial features of the latent code by a carefully designed dimension reduction component for distributions; 2. Modeling the latent distribution as a mixture of Gaussian low-rank inliers and full-rank outliers, where the testing only uses the inlier model; 3. Applying the Wasserstein-1 metric for regularization, instead of the Kullback-Leibler (KL) divergence; and 4. Using a least absolute deviation error for reconstruction. We establish both robustness to outliers and suitability to low-rank modeling of the Wasserstein metric as opposed to the KL divergence. We illustrate state-of-the-art results on standard benchmarks for novelty detection.
|
true |
Boost then Convolve: Gradient Boosting Meets Graph Neural Networks
GNN GBDT graphs tabular data heterogeneous data
Graph neural networks (GNNs) are powerful models that have been successful in various graph representation learning tasks. Whereas gradient boosted decision trees (GBDT) often outperform other machine learning methods when faced with heterogeneous tabular data. But what approach should be used for graphs with tabular node features? Previous GNN models have mostly focused on networks with homogeneous sparse features and, as we show, are suboptimal in the heterogeneous setting. In this work, we propose a novel architecture that trains GBDT and GNN jointly to get the best of both worlds: the GBDT model deals with heterogeneous features, while GNN accounts for the graph structure. Our model benefits from end-to-end optimization by allowing new trees to fit the gradient updates of GNN. With an extensive experimental comparison to the leading GBDT and GNN models, we demonstrate a significant increase in performance on a variety of graphs with tabular features. The code is available: https://github.com/nd7141/bgnn.
|
true |
Finding Structure and Causality in Linear Programs
linear programs structure causality graph learning
Linear Programs (LP) are celebrated widely, particularly so in machine learning where they have allowed for effectively solving probabilistic inference tasks or imposing structure on end-to-end learning systems. Their potential might seem depleted but we propose a foundational, causal perspective that reveals intriguing intra- and inter-structure relations for LP components. We conduct a systematic, empirical investigation on general-, shortest path- and energy system LPs.
|
false |
Adaptive Single-Pass Stochastic Gradient Descent in Input Sparsity Time
stochastic gradient descent streaming algorithm stochastic optimization
We study sampling algorithms for variance reduction methods for stochastic optimization. Although stochastic gradient descent (SGD) is widely used for large scale machine learning, it sometimes experiences slow convergence rates due to the high variance from uniform sampling. In this paper, we introduce an algorithm that approximately samples a gradient from the optimal distribution for a common finite-sum form with $n$ terms, while just making a single pass over the data, using input sparsity time, and $\tO{Td}$ space. Our algorithm can be implemented in big data models such as the streaming and distributed models. Moreover, we show that our algorithm can be generalized to approximately sample Hessians and thus provides variance reduction for second-order methods as well. We demonstrate the efficiency of our algorithm on large-scale datasets.
|
false |
Large Scale Image Completion via Co-Modulated Generative Adversarial Networks
GAN co-modulation image completion image-to-image translation
Co-modulated GANs link image-conditional GANs and unconditional modulated models to address large-scale image completion tasks. Co-modulation brings stochastic and conditional style representations together. To improve existing metrics for image completion, the proposed Paired/Unpaired Inception Discriminative Score (P-IDS/U-IDS) is robust to sampling size, captures subtle differences well, and correlates with human preferences. Experiments using co-modulated GANs lead to high quality and diverse results in free-form image completion and image-to-image translation tasks. We extend the findings by Zhao et al. by performing new image completion experiments to examine the biases of co-modulated GANs.
|
false |
One Vertex Attack on Graph Neural Networks-based Spatiotemporal Forecasting
adversarial attack graph neural networks spatiotemporal forecasting
Spatiotemporal forecasting plays an essential role in intelligent transportation systems (ITS) and numerous applications, such as route planning, navigation, and automatic driving. Deep Spatiotemporal Graph Neural Networks, which capture both spatial and temporal patterns, have achieved great success in traffic forecasting applications. Though Deep Neural Networks (DNNs) have been proven to be vulnerable to carefully designed perturbations in multiple domains like objection classification and graph classification, these adversarial works cannot be directly applied to spatiotemporal GNNs because of their causality and spatiotemporal mechanism. There is still a lack of studies on the vulnerability and robustness of spatiotemporal GNNs. Particularly, if spatiotemporal GNNs are vulnerable in real-world traffic applications, a hacker can easily cause serious traffic congestion and even a city-scale breakdown. To fill this gap, we design One Vertex Attack to break deep spatiotemporal GNNs by attacking a single one vertex. To achieve this, we apply the genetic algorithm with a universal attack method as the evaluation function to locate the weakest vertex; then perturbations are generated by solving an optimization problem with the inverse estimation. Empirical studies prove that perturbations in one vertex can be diffused into most of the graph when spatiotemporal GNNs are under One Vertex Attack.
|
false |
An Attention Free Transformer
Transformers attention efficient
We introduce Attention Free Transformer (AFT), an efficient variant of Transformers \citep{transformer} that eliminates the need for dot product attention. AFT offers great simplicity and efficiency compared with standard Transformers, where the multi-head attention operation is replaced with the composition of element-wise multiplications/divisions and global/local pooling. During training time, AFT has linear time and space complexity w.r.t. both the sequence length and feature dimension; in the autoregressive decoding mode, AFT has constant memory and time complexity per step. We show that, surprisingly, we are able to train AFT effectively on challenging benchmarks, and also to match or surpass the standard Transformer counterparts and other efficient variants. In particular, AFT achieves the state-of-the-art result on CIFAR10 autoregressive modeling with much reduced complexity, and also outperforms several efficient Transformer variants on Enwik8.
|
true |
DialoGraph: Incorporating Interpretable Strategy-Graph Networks into Negotiation Dialogues
negotiation dialogue graph neural networks interpretability structure
To successfully negotiate a deal, it is not enough to communicate fluently: pragmatic planning of persuasive negotiation strategies is essential. While modern dialogue agents excel at generating fluent sentences, they still lack pragmatic grounding and cannot reason strategically. We present DialoGraph, a negotiation system that incorporates pragmatic strategies in a negotiation dialogue using graph neural networks. DialoGraph explicitly incorporates dependencies between sequences of strategies to enable improved and interpretable prediction of next optimal strategies, given the dialogue context. Our graph-based method outperforms prior state-of-the-art negotiation models both in the accuracy of strategy/dialogue act prediction and in the quality of downstream dialogue response generation. We qualitatively show further benefits of learned strategy-graphs in providing explicit associations between effective negotiation strategies over the course of the dialogue, leading to interpretable and strategic dialogues.
|
true |
Implicit Neural Video Compression
implicit neural representation video compression optical flow model quantization
We propose a method to compress full-resolution video sequences with implicit neural representations. Each frame is represented as a neural network that maps coordinate positions to pixel values. We use a separate implicit network to modulate the coordinate inputs, which enables efficient motion compensation between frames. Together with a small residual network, this allows us to efficiently compress P-frames relative to the previous frame. We further lower the bitrate by storing the network weights with learned integer quantization. Our method offers several simplifications over established neural video codecs: it does not require the receiver to have access to a pretrained neural network, does not use expensive interpolation-based warping operations, and does not require a separate training dataset.
|
false |
Small Input Noise is Enough to Defend Against Query-based Black-box Attacks
Gaussian noise input noise adversarial defense black-box attack adversarial attack query-based attack
While deep neural networks show unprecedented performance in various tasks, the vulnerability to adversarial examples hinders their deployment in safety-critical systems. Many studies have shown that attacks are also possible even in a black-box setting where an adversary cannot access the target model's internal information. Most black-box attacks are based on queries, each of which obtains the target model's output for an input, and many recent studies focus on reducing the number of required queries. In this paper, we pay attention to an implicit assumption of these attacks that the target model's output exactly corresponds to the query input. If some randomness is introduced into the model to break this assumption, query-based attacks may have tremendous difficulty in both gradient estimation and local search, which are the core of their attack process. From this motivation, we observe even a small additive input noise can neutralize most query-based attacks and name this simple yet effective approach Small Noise Defense (SND). We analyze how SND can defend against query-based black-box attacks and demonstrate its effectiveness against eight different state-of-the-art attacks with CIFAR-10 and ImageNet datasets. Even with strong defense ability, SND almost maintains the original clean accuracy and computational speed. SND is readily applicable to pre-trained models by adding only one line of code at the inference stage, so we hope that it will be used as a baseline of defense against query-based black-box attacks in the future.
|
false |
Learning to Plan Optimistically: Uncertainty-Guided Deep Exploration via Latent Model Ensembles
Model-Based Reinforcement Learning Deep Exploration Continuous Visual Control UCB Latent Space Ensembling
Learning complex behaviors through interaction requires coordinated long-term planning. Random exploration and novelty search lack task-centric guidance and waste effort on non-informative interactions. Instead, decision making should target samples with the potential to optimize performance far into the future, while only reducing uncertainty where conducive to this objective. This paper presents latent optimistic value exploration (LOVE), a strategy that enables deep exploration through optimism in the face of uncertain long-term rewards. We combine finite-horizon rollouts from a latent model with value function estimates to predict infinite-horizon returns and recover associated uncertainty through ensembling. Policy training then proceeds on an upper confidence bound (UCB) objective to identify and select the interactions most promising to improve long-term performance. We apply LOVE to continuous visual control tasks and demonstrate improved sample complexity on a selection of benchmarking tasks.
|
false |
Consensus Clustering with Unsupervised Representation Learning
Clustering Ensemble Learning Representation Learning
Recent advances in deep clustering and unsupervised representation learning are based on the idea that different views of an input image (generated through data augmentation techniques) must either be closer in the representation space, or have a similar cluster assignment. In this work, we leverage this idea together with ensemble learning to perform clustering and representation learning. Ensemble learning is widely used in the supervised learning setting but has not yet been practical in deep clustering. Previous works on ensemble learning for clustering neither work on the feature space nor learn features. We propose a novel ensemble learning algorithm dubbed Consensus Clustering with Unsupervised Representation Learning (ConCURL) which learns representations by creating a consensus on multiple clustering outputs. Specifically, we generate a cluster ensemble using random transformations on the embedding space, and define a consensus loss function that measures the disagreement among the constituents of the ensemble. Thus, diverse ensembles minimize this loss function in a synergistic way, which leads to better representations that work with all cluster ensemble constituents. Our proposed method ConCURL is easy to implement and integrate into any representation learning or deep clustering block. ConCURL outperforms all state of the art methods on various computer vision datasets. Specifically, we beat the closest state of the art method by 5.9 percent on the ImageNet-10 dataset, and by 18 percent on the ImageNet-Dogs dataset in terms of clustering accuracy. We further shed some light on the under-studied overfitting issue in clustering and show that our method does not overfit as much as existing methods, and thereby generalizes better for new data samples.
|
false |
Performant LLM Agentic Framework for Conversational AI
agentic agentic ai conversational ai machine learning workflow navigation ai automation agent performant latency
The rise of Agentic applications and automation in the Voice AI industry has led to an increased reliance on Large Language Models (LLMs) to navigate graph-based logic workflows composed of nodes and edges. However, existing methods face challenges such as alignment errors in complex workflows and hallucinations caused by excessive context size. To address these limitations, we introduce the Performant Agentic Framework (PAF), a novel system that assists LLMs in selecting appropriate nodes and executing actions in order when traversing complex graphs. PAF combines LLM-based reasoning with a mathematically grounded vector scoring mechanism, achieving both higher accuracy and reduced latency. Our approach dynamically balances strict adherence to predefined paths with flexible node jumps to handle various user inputs efficiently. Experiments demonstrate that PAF significantly outperforms baseline methods, paving the way for scalable, real-time Conversational AI systems in complex business environments.
|
true |
Minimal-Entropy Correlation Alignment for Unsupervised Deep Domain Adaptation
unsupervised domain adaptation entropy minimization image classification deep transfer learning
In this work, we face the problem of unsupervised domain adaptation with a novel deep learning approach which leverages our finding that entropy minimization is induced by the optimal alignment of second order statistics between source and target domains. We formally demonstrate this hypothesis and, aiming at achieving an optimal alignment in practical cases, we adopt a more principled strategy which, differently from the current Euclidean approaches, deploys alignment along geodesics. Our pipeline can be implemented by adding to the standard classification loss (on the labeled source domain), a source-to-target regularizer that is weighted in an unsupervised and data-driven fashion. We provide extensive experiments to assess the superiority of our framework on standard domain and modality adaptation benchmarks.
|
false |
A Bayesian-Symbolic Approach to Learning and Reasoning for Intuitive Physics
physics learning symbolic regression intuitive physics
Humans are capable of reasoning about physical phenomena by inferring laws of physics from a very limited set of observations. The inferred laws can potentially depend on unobserved properties, such as mass, texture, charge, etc. This sample-efficient physical reasoning is considered a core domain of human common-sense knowledge and hints at the existence of a physics engine in the head. In this paper, we propose a Bayesian symbolic framework for learning sample-efficient models of physical reasoning and prediction, which are of special interests in the field of intuitive physics. In our framework, the environment is represented by a top-down generative model with a collection of entities with some known and unknown properties as latent variables to capture uncertainty. The physics engine depends on physical laws which are modeled as interpretable symbolic expressions and are assumed to be functions of the latent properties of the entities interacting under simple Newtonian physics. As such, learning the laws is then reduced to symbolic regression and Bayesian inference methods are used to obtain the distribution of unobserved properties. These inference and regression steps are performed in an iterative manner following the expectation–maximization algorithm to infer the unknown properties and use them to learn the laws from a very small set of observations. We demonstrate that on three physics learning tasks that compared to the existing methods of learning physics, our proposed framework is more data-efficient, accurate and makes joint reasoning and learning possible.
|
false |
DISE: Dynamic Integrator Selection to Minimize Forward Pass Time in Neural ODEs
Neural ODE DOPRI
Neural ordinary differential equations (Neural ODEs) are appreciated for their ability to significantly reduce the number of parameters when constructing a neural network. On the other hand, they are sometimes blamed for their long forward-pass inference time, which is incurred by solving integral problems. To improve the model accuracy, they rely on advanced solvers, such as the Dormand--Prince (DOPRI) method. To solve an integral problem, however, it requires at least tens (or sometimes thousands) of steps in many Neural ODE experiments. In this work, we propose to i) directly regularize the step size of DOPRI to make the forward-pass faster and ii) dynamically choose a simpler integrator than DOPRI for a carefully selected subset of input. Because it is not the case that every input requires the advanced integrator, we design an auxiliary neural network to choose an appropriate integrator given input to decrease the overall inference time without significantly sacrificing accuracy. We consider the Euler method, the fourth-order Runge--Kutta (RK4) method, and DOPRI as selection candidates. We found that 10-30% of cases can be solved with simple integrators in our experiments. Therefore, the overall number of functional evaluations (NFE) decreases up to 78% with improved accuracy.
|
false |
Double Q-learning: New Analysis and Sharper Finite-time Bound
Double Q-learning Finite-time analysis Convergence rate Stochastic approximation
Double Q-learning \citep{hasselt2010double} has gained significant success in practice due to its effectiveness in overcoming the overestimation issue of Q-learning. However, theoretical understanding of double Q-learning is rather limited and the only existing finite-time analysis was recently established in \citet{xiong2020double} under a polynomial learning rate. This paper analyzes the more challenging case with a rescaled linear/constant learning rate for which the previous method does not appear to be applicable. We develop new analytical tools that achieve an order-level better finite-time convergence rate than the previously established result. Specifically, we show that synchronous double Q-learning attains an $\epsilon$-accurate global optimum with a time complexity of $\Omega\left(\frac{\ln D}{(1-\gamma)^7\epsilon^2} \right)$, and the asynchronous algorithm attains a time complexity of $\tilde{\Omega}\left(\frac{L}{(1-\gamma)^7\epsilon^2} \right)$, where $D$ is the cardinality of the state-action space, $\gamma$ is the discount factor, and $L$ is a parameter related to the sampling strategy for asynchronous double Q-learning. These results improve the order-level dependence of the convergence rate on all major parameters $(\epsilon,1-\gamma, D, L)$ provided in \citet{xiong2020double}. The new analysis in this paper presents a more direct and succinct approach for characterizing the finite-time convergence rate of double Q-learning.
|
true |
Physics-informed neural networks for sampling
sampling partial differential equations physics-informed neural networks
We present a framework to sample from high-dimensional unnormalized densities using physics-informed neural networks (PINNs). For various computational science tasks, it is essential to draw samples from a target distribution where the density is known up to a normalizing constant. Without access to any training samples, existing methods based on normalizing flows and diffusion models rely on the simulation of (stochastic) differential equations for training and suffer from mode collapse. Our approach circumvents these issues by solving the underlying continuity and Fokker-Planck equations using PINNs. Motivated by optimal transport and Schrödinger bridges, we further incorporate regularizers based on Hamilton-Jacobi-Bellman equations. Through evaluations on several benchmarks, we demonstrate that our approach can mitigate mode collapse and significantly outperform various baselines.
|
false |
GeDi: Generative Discriminator Guided Sequence Generation
Language modeling controllable generation decoding schemes auto-regressive models language modeling safety
While large-scale language models (LMs) are able to imitate the distribution of natural language well enough to generate realistic text, it is difficult to control which regions of the distribution they generate. This is especially problematic because datasets used for training large LMs usually contain significant toxicity, hate, bias, and negativity. We propose GeDi as an efficient method for using smaller LMs as generative discriminators to guide generation from large LMs to make them safer and more controllable. GeDi guides generation at each step by computing classification probabilities for all possible next tokens via Bayes rule by normalizing over two class-conditional distributions; one conditioned on the desired attribute, or control code, and another conditioned on the undesired attribute, or anti control code. We find that GeDi gives controllability on par with or better than the state of the art method in a variety of settings, while also achieving generation speeds more than $30$ times faster. Additionally, training GeDi on only three topics allows us to controllably generate new topics zero-shot from just a keyword. Lastly, we show that GeDi can make GPT-2 and GPT-3 significantly less toxic without sacrificing on linguistic fluency, making it by far the most practical existing method for detoxifying large language models while maintaining a fast generation speed.
|
true |
Selective Classification Can Magnify Disparities Across Groups
selective classification group disparities log-concavity robustness
Selective classification, in which models can abstain on uncertain predictions, is a natural approach to improving accuracy in settings where errors are costly but abstentions are manageable. In this paper, we find that while selective classification can improve average accuracies, it can simultaneously magnify existing accuracy disparities between various groups within a population, especially in the presence of spurious correlations. We observe this behavior consistently across five vision and NLP datasets. Surprisingly, increasing abstentions can even decrease accuracies on some groups. To better understand this phenomenon, we study the margin distribution, which captures the model’s confidences over all predictions. For symmetric margin distributions, we prove that whether selective classification monotonically improves or worsens accuracy is fully determined by the accuracy at full coverage (i.e., without any abstentions) and whether the distribution satisfies a property we call left-log-concavity. Our analysis also shows that selective classification tends to magnify full-coverage accuracy disparities. Motivated by our analysis, we train distributionally-robust models that achieve similar full-coverage accuracies across groups and show that selective classification uniformly improves each group on these models. Altogether, our results suggest that selective classification should be used with care and underscore the importance of training models to perform equally well across groups at full coverage.
|
false |
GL-Disen: Global-Local disentanglement for unsupervised learning of graph-level representations
Unsupervised Graph Representations Disentanglement Learning GNN Unsupervised Learning
Graph-level representation learning plays a crucial role in a variety of tasks such as molecular property prediction and community analysis. Currently, several models based on mutual information maximization have shown strong performance on the task of unsupervised graph representation learning. In this paper, instead, we consider a disentanglement approach to learn graph-level representations in the unsupervised setting. Our work is the first to study disentanglement learning for graph-level representations. Our key observation is that the formation of many real-world graphs is a complex process with global and local generative factors. We hypothesize that disentangled representations which capture these global and local generative factors into independent latent units can be highly beneficial. Specifically, for graph-level representation learning, our disentanglement approach can alleviate distraction due to local variations of individual nodes or individual local neighbourhoods. We propose a VAE based learning algorithm to disentangle the global graph-level information, which is common across the entire graph, and local patch-level information, which varies across individual patches (the local subgraphs centered around the nodes). Through extensive experiments and analysis, we show that our method achieves the state-of-the-art performance on the task of unsupervised graph representation learning.
|
false |
Deep Q-Learning with Low Switching Cost
deep Q-network DQN switching cost deep Q-learning
We initiate the study on deep reinforcement learning problems that require low switching cost, i.e., small number of policy switches during training. Such a requirement is ubiquitous in many applications, such as medical domains, recommendation systems, education, robotics, dialogue agents, etc, where the deployed policy that actually interacts with the environment cannot change frequently. Our paper investigates different policy switching criteria based on deep Q-networks and further proposes an adaptive approach based on the feature distance between the deployed Q-network and the underlying learning Q-network. Through extensive experiments on a medical treatment environment and a collection of the Atari games, we find our feature-switching criterion substantially decreases the switching cost while maintains a similar sample efficiency to the case without the low-switching-cost constraint. We also complement this empirical finding with a theoretical justification from a representation learning perspective.
|
true |
The Annotated S4
annotated implementation series papers models sequence modeling hippo recurrent memory optimal polynomial projections
An annotated implementation of a series of papers developing state-space models for very long-term sequence modeling. Covers "HiPPO: Recurrent Memory with Optimal Polynomial Projections" and ends with "Efficiently Modeling Long Sequences with Structured State Spaces".
|
true |
Property Controllable Variational Autoencoder via Invertible Mutual Dependence
deep generative models interpretable latent representation disentangled representation learning
Deep generative models have made important progress towards modeling complex, high dimensional data via learning latent representations. Their usefulness is nevertheless often limited by a lack of control over the generative process or a poor understanding of the latent representation. To overcome these issues, attention is now focused on discovering latent variables correlated to the data properties and ways to manipulate these properties. This paper presents the new Property controllable VAE (PCVAE), where a new Bayesian model is proposed to inductively bias the latent representation using explicit data properties via novel group-wise and property-wise disentanglement. Each data property corresponds seamlessly to a latent variable, by innovatively enforcing invertible mutual dependence between them. This allows us to move along the learned latent dimensions to control specific properties of the generated data with great precision. Quantitative and qualitative evaluations confirm that the PCVAE outperforms the existing models by up to 28% in capturing and 65% in manipulating the desired properties.
|
true |
Mixture of Neural Operators: Incorporating Historical Information for Longer Rollouts
Neural PDE Solvers Neural Operators Autoregressive Methods
Traditional numerical solvers for time-dependent partial differential equations (PDEs) notoriously require high computational resources and necessitate recomputation when faced with new problem parameters. In recent years, neural surrogates have shown great potential to overcome these limitations. However, it has been paradoxically observed that incorporating historical information into neural surrogates worsens their rollout performance. Drawing inspiration from multistep methods that use historical information from previous steps to obtain higher-order accuracy, we introduce the Mixture of Neural Operators (MoNO) framework; a collection of neural operators, each dedicated to processing information from a distinct previous step. We validate MoNO on the Kuramoto-Sivashinsky equation, demonstrating enhanced accuracy and stability of longer rollouts, greatly outperforming neural operators that discard historical information.
|
false |
Learning-Augmented Sketches for Hessians
sketches hessians sketching considerable speedups second order optimization number works hessian iteration specific
We study learning-based sketching for Hessians, which is known to provide considerable speedups to second order optimization. A number of works have shown how to sketch or subsample the Hessian to speed up each iteration, but such sketches are usually specific to the matrix at hand, rather than being learned from a distribution. We extend such schemes to learned sketches, where we learn different potentially different sketches for the different iterations, and show empirically that learned sketches, compared with their "non-learned" counterparts, improve the approximation accuracy for a large number of important problems, including LASSO, SVM, and matrix estimation with nuclear norm constraints.
|
false |
Learning Robust Models using the Principle of Independent Causal Mechanisms
Causal Discovery Principle of Independent Causal Mechanisms Normalizing Flows Domain Generalization
Standard supervised learning breaks down under data distribution shift. However, the principle of independent causal mechanisms (ICM, Peters et al. (2017)) can turn this weakness into an opportunity: one can take advantage of distribution shift between different environments during training in order to obtain more robust models. We propose a new gradient-based learning framework whose objective function is derived from the ICM principle. We show theoretically and experimentally that neural networks trained in this framework focus on relations remaining invariant across environments and ignore unstable ones. Moreover, we prove that the recovered stable relations correspond to the true causal mechanisms under certain conditions. In both regression and classification, the resulting models generalize well to unseen scenarios where traditionally trained models fail.
|
true |
UNCONDITIONAL IMAGE-TEXT PAIR GENERATION WITH MULTIMODAL CROSS QUANTIZER
Multimodal Unconditional generation Vector quantization Joint representation
Though deep generative models have gained a lot of attention, most of the existing works are designed for the unimodal generation task. In this paper, we explore a new method for unconditional image-text pair generation. We propose MXQ-VAE, a vector quantization method for multimodal image-text representation. MXQ-VAE accepts a paired image and text as input, and learns a joint quantized representation space, so that the image-text pair can be converted to a sequence of unified indices. Then we can use autoregressive generative models to model the joint image-text representation, and even perform unconditional image-text pair generation. Extensive experimental results demonstrate that our approach effectively generates semantically consistent image-text pair and also enhances meaningful alignment between image and text.
|
true |
Connecting the Dots Between MLE and RL for Sequence Generation
sequence generation maximum likelihood learning reinforcement learning policy optimization text generation reward augmented maximum likelihood exposure bias
Sequence generation models such as recurrent networks can be trained with a diverse set of learning algorithms. For example, maximum likelihood learning is simple and efficient, yet suffers from the exposure bias problem. Reinforcement learning like policy gradient addresses the problem but can have prohibitively poor exploration efficiency. A variety of other algorithms such as RAML, SPG, and data noising, have also been developed in different perspectives. This paper establishes a formal connection between these algorithms. We present a generalized entropy regularized policy optimization formulation, and show that the apparently divergent algorithms can all be reformulated as special instances of the framework, with the only difference being the configurations of reward function and a couple of hyperparameters. The unified interpretation offers a systematic view of the varying properties of exploration and learning efficiency. Besides, based on the framework, we present a new algorithm that dynamically interpolates among the existing algorithms for improved learning. Experiments on machine translation and text summarization demonstrate the superiority of the proposed algorithm.
|
true |
Cross-Linked Variational Autoencoders for Generalized Zero-Shot Learning
generalized zero-shot learning zero-shot learning few-shot learning image classification
Most approaches in generalized zero-shot learning rely on cross-modal mapping between an image feature space and a class embedding space or on generating artificial image features. However, learning a shared cross-modal embedding by aligning the latent spaces of modality-specific autoencoders is shown to be promising in (generalized) zero-shot learning. While following the same direction, we also take artificial feature generation one step further and propose a model where a shared latent space of image features and class embeddings is learned by aligned variational autoencoders, for the purpose of generating latent features to train a softmax classifier. We evaluate our learned latent features on conventional benchmark datasets and establish a new state of the art on generalized zero-shot as well as on few-shot learning. Moreover, our results on ImageNet with various zero-shot splits show that our latent features generalize well in large-scale settings.
|
false |
Benchmarking Unsupervised Object Representations for Video Sequences
Unsupervised learning object-centric representations benchmark tracking
Perceiving the world in terms of objects and tracking them through time is a crucial prerequisite for reasoning and scene understanding. Recently, several methods have been proposed for unsupervised learning of object-centric representations. However, since these models have been evaluated with respect to different downstream tasks, it remains unclear how they compare in terms of basic perceptual abilities such as detection, figure-ground segmentation and tracking of individual objects. To close this gap, we design a benchmark with three datasets of varying complexity and seven additional test sets which feature challenging tracking scenarios relevant for natural videos. Using this benchmark, we compare the perceptual abilities of four unsupervised object-centric learning approaches: ViMON, a video-extension of MONet, based on a recurrent spatial attention mechanism, OP3, which exploits clustering via spatial mixture models, as well as TBA and SCALOR, which use an explicit factorization via spatial transformers. Our results suggest that architectures with unconstrained latent representations and full-image object masks such as ViMON and OP3 are able to learn more powerful representations in terms of object detection, segmentation and tracking than the explicitly parameterized spatial transformer based architecture of TBA and SCALOR. We also observe that none of the methods are able to gracefully handle the most challenging tracking scenarios despite their synthetic nature, suggesting that our benchmark may provide fruitful guidance towards learning more robust object-centric video representations.
|
false |
Certified Robustness of Nearest Neighbors against Data Poisoning Attacks
attacks nearest neighbors data knn certified defenses rnn certified robustness robustness attacks data
Data poisoning attacks aim to corrupt a machine learning model via modifying, adding, and/or removing some carefully selected training examples, such that the corrupted model predicts any or attacker-chosen incorrect labels for testing examples. The key idea of state-of-the-art certified defenses against data poisoning attacks is to create a \emph{majority vote} mechanism to predict the label of a testing example. Moreover, each voter is a base classifier trained on a subset of the training dataset. Nearest neighbor algorithms such as $k$ nearest neighbors (kNN) and radius nearest neighbors (rNN) have intrinsic majority vote mechanisms. In this work, we show that the intrinsic majority vote mechanisms in kNN and rNN already provide certified robustness guarantees against general data poisoning attacks. Moreover, our empirical evaluation results on MNIST and CIFAR10 show that the intrinsic certified robustness guarantees of kNN and rNN outperform those provided by state-of-the-art certified defenses.
|
false |
Incorporating Bayesian approaches in Deep Learning Research
bayesian-statistics machine-learning
The blog will take the form of a survey paper as we summarize the proceedings from the [Symposium on Approximate Bayesian Inference](http://approximateinference.org/), during [NeurIPS](https://nips.cc/), 2019 in Vancouver. Rightly so, I focused on the Bayesian workshop as it emphasized the fundamentals, rather than bleeding-edge results. The knowledge that can lead to progress results from understanding how things work on a foundational level.
|
false |
Brain-like approaches to unsupervised learning of hidden representations - a comparative study
neural networks bio-inspired brain-like unsupervised learning structural plasticity
Unsupervised learning of hidden representations has been one of the most vibrant research directions in machine learning in recent years. In this work we study the brain-like Bayesian Confidence Propagating Neural Network (BCPNN) model, recently extended to extract sparse distributed high-dimensional representations. The saliency and separability of the hidden representations when trained on MNIST dataset is studied using an external linear classifier and compared with other unsupervised learning methods that include restricted Boltzmann machines and autoencoders.
|
false |
What is the chance of being so unfair?
fairness ranking
Fairness has often been seen as an ethical concern that needs to be considered at some cost on the utility. In contrast, in this work, we formulate fairness, and especially fairness in ranking, as a way to avoid unjust biases and provide a more accurate ranking that results in improvement on the actual unbiased utility. With this in mind, we design a fairness measure that, instead of blindly forcing some approximate equality constraint, checks if the outcome is plausible in a just world. Our fairness measure asks a simple and fundamental statistical question: "What is the chance of observing this outcome in an unbiased world?". If the chance is high enough, the outcome is fair. We provide a dynamic programming algorithm that, given a ranking calculates our fairness measure. Secondly, given a sequence of potentially biased scores, along with the sensitive feature, we provide a fair ranking algorithm based on our fairness measure. Finally, we run some experiments to understand the behavior of our ranking algorithm against other fundamental algorithms.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.