Spaces:
Running
Running
File size: 12,874 Bytes
d9e62f5 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 |
{
"cells": [
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"import requests\n",
"url = \"http://export.arxiv.org/api/query?search_query=all:deep+learning&start=0&max_results=5\"\n",
"response = requests.get(url)"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [],
"source": [
"import xml.etree.ElementTree as ET\n",
"\n",
"# Parse the XML response\n",
"root = ET.fromstring(response.text)\n",
"\n",
"# Extract titles and abstracts\n",
"abstracts = [] # List to store abstracts\n",
"titles = [] # List to store titles\n",
"\n",
"for entry in root.findall(\"{http://www.w3.org/2005/Atom}entry\"):\n",
" title = entry.find(\"{http://www.w3.org/2005/Atom}title\").text.strip()\n",
" abstract = entry.find(\"{http://www.w3.org/2005/Atom}summary\").text.strip()\n",
"\n",
" titles.append(title)\n",
" abstracts.append(abstract)"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/Users/nitin/Documents/AI Agent project/.venv/lib/python3.12/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
" from .autonotebook import tqdm as notebook_tqdm\n",
"Device set to use mps:0\n"
]
}
],
"source": [
"from transformers import pipeline\n",
"# Initialize summarization pipeline\n",
"summarizer = pipeline(\"summarization\", model=\"facebook/bart-large-cnn\")"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"Asking to truncate to max_length but no maximum length is provided and the model has no predefined maximum length. Default to no truncation.\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Title: Opening the black box of deep learning\n",
"Summary: The great success of deep learning shows that its technology contains aprofound truth. Understanding its internal mechanism not only has important implications for the development of its technology and effective application in various fields. At present, most of the theoretical research on\n",
"--------------------------------------------------------------------------------\n",
"Title: Concept-Oriented Deep Learning\n",
"Summary: Concepts are the foundation of human deep learning, understanding, and knowledge integration and transfer. We propose concept-oriented deep learning(CODL) which extends (machine) deep learning with concept representations. CODL addresses some\n",
"--------------------------------------------------------------------------------\n",
"Title: Deep learning research landscape & roadmap in a nutshell: past, present\n",
" and future -- Towards deep cortical learning\n",
"Summary: The past, present and future of deep learning is presented in this work. We predict that deep cortical learning will be the convergence of deeplearning & cortical learning which builds an artificial cortical column.\n",
"--------------------------------------------------------------------------------\n",
"Title: A First Look at Deep Learning Apps on Smartphones\n",
"Summary: First empirical study on 16,500 popular Android apps. Demystifies how smartphone apps exploit deep learning in the wild. Findings paint promising picture of deep learning for smartphones.\n",
"--------------------------------------------------------------------------------\n",
"Title: Geometrization of deep networks for the interpretability of deep\n",
" learning systems\n",
"Summary: Geometrization is a bridge to connect physics, geometry, deep networkand quantum computation. This may result in a new scheme to reveal the rule of the physical world. It may also help to solve theinterpretability problem of deep\n",
"--------------------------------------------------------------------------------\n"
]
}
],
"source": [
"# Summarize abstracts\n",
"summaries = [summarizer(abs, max_length=50, min_length=25, truncation=True)[0][\"summary_text\"] for abs in abstracts]\n",
"\n",
"# Print results\n",
"for title, summary in zip(titles, summaries):\n",
" print(f\"Title: {title}\\nSummary: {summary}\\n\" + \"-\"*80)"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/var/folders/cv/01jc2kcx1455j48m7hhwcvqw0000gn/T/ipykernel_17287/3038899353.py:4: LangChainDeprecationWarning: The class `HuggingFaceEmbeddings` was deprecated in LangChain 0.2.2 and will be removed in 1.0. An updated version of the class exists in the :class:`~langchain-huggingface package and should be used instead. To use it run `pip install -U :class:`~langchain-huggingface` and import as `from :class:`~langchain_huggingface import HuggingFaceEmbeddings``.\n",
" embedding_model = HuggingFaceEmbeddings(model_name=\"sentence-transformers/all-MiniLM-L6-v2\")\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Top 1 Result:\n",
"We are in the dawn of deep learning explosion for smartphones. To bridge the\n",
"gap between research and practice, we present the first empirical study on\n",
"16,500 the most popular Android apps, demystifying how smartphone apps exploit\n",
"deep learning in the wild. To this end, we build a new static tool that\n",
"dissects apps and analyzes their deep learning functions. Our study answers\n",
"threefold questions: what are the early adopter apps of deep learning, what do\n",
"they use deep learning for, and how do their deep learning models look like.\n",
"Our study has strong implications for app developers, smartphone vendors, and\n",
"deep learning R\\&D. On one hand, our findings paint a promising picture of deep\n",
"learning for smartphones, showing the prosperity of mobile deep learning\n",
"frameworks as well as the prosperity of apps building their cores atop deep\n",
"learning. On the other hand, our findings urge optimizations on deep learning\n",
"models deployed on smartphones, the protection of these models, and validation\n",
"of research ideas on these models.\n",
"--------------------------------------------------------------------------------\n",
"Top 2 Result:\n",
"The past, present and future of deep learning is presented in this work.\n",
"Given this landscape & roadmap, we predict that deep cortical learning will be\n",
"the convergence of deep learning & cortical learning which builds an artificial\n",
"cortical column ultimately.\n",
"--------------------------------------------------------------------------------\n"
]
}
],
"source": [
"from langchain.vectorstores import FAISS\n",
"from langchain.embeddings import HuggingFaceEmbeddings\n",
"# Initialize HuggingFace Embeddings\n",
"embedding_model = HuggingFaceEmbeddings(model_name=\"sentence-transformers/all-MiniLM-L6-v2\")\n",
"\n",
"# Create FAISS Vector Store\n",
"vector_store = FAISS.from_texts(abstracts, embedding_model)\n",
"\n",
"# Perform a similarity search (RAG Retrieval Step)\n",
"query = \"best deep learning advancements\"\n",
"top_docs = vector_store.similarity_search(query, k=2)\n",
"\n",
"# Display Retrieved Documents\n",
"for i, doc in enumerate(top_docs):\n",
" print(f\"Top {i+1} Result:\\n{doc.page_content}\\n\" + \"-\"*80)"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"Device set to use mps:0\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Agent: Time to flex my research muscles!\n",
"Top papers: ['First empirical study on 16,500 popular Android apps. Demystifies how smartphone apps exploit deep learning in the wild. Findings', 'The past, present and future of deep learning is presented in this work. We predict that deep cortical learning will be the convergence of deep']\n",
"Best pick: First empirical study on 16,500 popular Android apps. Demystifies how smartphone apps exploit deep learning in the wild. Findings\n"
]
}
],
"source": [
"# Prerequisites: pip install langchain faiss-cpu transformers requests\n",
"import requests\n",
"from langchain.vectorstores import FAISS\n",
"from langchain.embeddings import HuggingFaceEmbeddings\n",
"from transformers import pipeline\n",
"\n",
"# Step 1: Define the Agent class (our Research Buddy)\n",
"class ResearchBuddyAgent:\n",
" def __init__(self):\n",
" # Load embeddings for RAG\n",
" self.embeddings = HuggingFaceEmbeddings(model_name=\"all-MiniLM-L6-v2\")\n",
" # Load summarizer (small model for speed)\n",
" self.summarizer = pipeline(\"summarization\", model=\"facebook/bart-large-cnn\")\n",
" self.papers = [] # Store fetched papers\n",
"\n",
" # Step 2: Fetch papers (simple arXiv API call)\n",
" def fetch_papers(self, topic):\n",
" url = f\"http://export.arxiv.org/api/query?search_query=all:{topic}&start=0&max_results=5\"\n",
" response = requests.get(url)\n",
" # Parse the XML response\n",
" root = ET.fromstring(response.text)\n",
"\n",
" # Extract titles and abstracts\n",
" abstracts = [] # List to store abstracts\n",
" titles = [] # List to store titles\n",
"\n",
" for entry in root.findall(\"{http://www.w3.org/2005/Atom}entry\"):\n",
" title = entry.find(\"{http://www.w3.org/2005/Atom}title\").text.strip()\n",
" abstract = entry.find(\"{http://www.w3.org/2005/Atom}summary\").text.strip()\n",
"\n",
" titles.append(title)\n",
" abstracts.append(abstract)\n",
"\n",
" # Step 3: Build RAG vector store and retrieve\n",
" def retrieve_relevant(self, query):\n",
" # Create FAISS vector store from papers\n",
" vector_store = FAISS.from_texts(abstracts, embedding_model)\n",
" # Search for top matches\n",
" top_docs = vector_store.similarity_search(query, k=2)\n",
" return [doc.page_content for doc in top_docs] # Extract text\n",
"\n",
" # Step 4: Summarize papers\n",
" def summarize_papers(self, papers):\n",
" summaries = []\n",
" for paper in papers:\n",
" summary = self.summarizer(paper, max_length=30, min_length=15)[0][\"summary_text\"]\n",
" summaries.append(summary)\n",
" return summaries\n",
"\n",
" # Step 5: Decide the \"best\" (simple rule: shortest summary wins)\n",
" def pick_best(self, summaries):\n",
" best = min(summaries, key=len) # Lazy rule for demo\n",
" return best\n",
"\n",
" # Step 6: Main agent flow (plan and execute)\n",
" def run(self, topic, query):\n",
" print(\"Agent: Time to flex my research muscles!\")\n",
" # Plan: Fetch → Retrieve → Summarize → Pick\n",
" self.fetch_papers(topic)\n",
" relevant_papers = self.retrieve_relevant(query)\n",
" summaries = self.summarize_papers(relevant_papers)\n",
" best_summary = self.pick_best(summaries)\n",
" print(f\"Top papers: {summaries}\")\n",
" print(f\"Best pick: {best_summary}\")\n",
"\n",
"# Run the agent\n",
"if __name__ == \"__main__\":\n",
" agent = ResearchBuddyAgent()\n",
" agent.run(topic=\"deep learning\", query=\"best deep learning advancements\")"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": ".venv",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.12.2"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
|