arxiv OpenAI Codex Dominates July 2021 arXiv Discussion OpenAI's Codex model tops the list of most-discussed arXiv papers from July 2021. The code-generating system shows significant improvements over previous language models.
arxiv OpenAI Codex Dominates July 2021 arXiv Hot Papers List OpenAI's Codex, a GPT language model fine-tuned on code, tops the list of most tweeted arXiv papers from July 2021. The model significantly outperforms its predecessors in writing functional Python code.
openai OpenAI and GitHub Launch AI Copilot to Write and Complete Code GitHub and OpenAI introduce Copilot, an AI assistant that writes entire functions based on simple comments. The tool relies on a specialized GPT-3 model trained on billions of lines of public code.
openai OpenAI Details Codex, the Deep Learning Model Powering GitHub Copilot OpenAI publishes a new paper explaining how it adapted GPT-3 into Codex, the specialized deep learning model behind GitHub's Copilot coding assistant. By switching to supervised learning and adding code to the training data, researchers significantly boost the model's programming capabilities.
salesforce Salesforce Open-Sources CodeT5 to Improve AI Code Generation Salesforce researchers release CodeT5, an open-source AI system that understands and generates code by incorporating structural programming knowledge instead of relying solely on natural language processing techniques.
stanford Stanford Launches Center to Study AI Foundation Models Stanford HAI introduces the Center for Research on Foundation Models to unite over 175 researchers in studying the technology and societal impact of systems like GPT-3. The initiative focuses on democratizing access to these powerful AI models.
openai OpenAI Codex Outperforms Standard Language Models in Python Code Generation Researchers introduce Codex, a GPT model fine-tuned on GitHub code that successfully solves nearly 29 percent of Python programming challenges. A distinct version of this model currently powers the GitHub Copilot coding assistant.
codex OpenAI Codex Tackles Python Generation Using Repeated Sampling A new GPT language model called Codex demonstrates strong Python writing capabilities by generating solutions from docstrings. The model achieves a high success rate through a strategy of repeated sampling.
deepmind DeepMind Database Provides Predicted Structures for 200 Million Proteins DeepMind expands its AlphaFold database to include predicted structures for over 200 million known proteins, offering a massive free resource for global scientific research.
github GitHub Copilot Raises Security and Legal Concerns in Early Testing Early testers of GitHub's AI coding assistant Copilot encounter issues ranging from synthesized API keys to copyright and ethics questions.
openai GitHub Copilot Reveals Unexpected Shift in AI Language Model Applications OpenAI and GitHub release Copilot, an AI pair programmer built on a specialized version of GPT-3 called Codex. The tool highlights how general-purpose language models often find their most successful use in specific, structured domains like programming.
openai GitHub Copilot Reveals Unexpected Shift in AI Language Model Applications GitHub and OpenAI release Copilot, an AI pair programmer built on a specialized version of GPT-3 called Codex. The tool highlights how general-purpose language models are finding unexpected success in structured domains like programming.
arxiv Top Trending AI Research Papers on arXiv Highlighted for June 2021 A new roundup showcases the most tweeted artificial intelligence research papers uploaded to arXiv during June 2021. Notable highlights include an open-source system for detecting cactus diseases and a new method for efficiently fine-tuning large language models.
deepmind DeepMind AGI Research Highlights Widespread Unreadiness for Ethical AI A groundbreaking DeepMind paper argues that reinforcement learning accelerates the path to artificial general intelligence, raising urgent ethical concerns.
diffusion models New Diffusion Models Surpass GANs in High-Resolution Image Synthesis Researchers demonstrate that diffusion models achieve state-of-the-art image sample quality, outperforming GANs by using improved architectures and classifier guidance. The models achieve record-breaking scores across various ImageNet resolutions while maintaining excellent distribution coverage.