Reinforcement Learning

concepts · 53 notes linked

Related: Deepmind · Large Language Models · Openai · AI Agents · Deep Learning · Machine Learning · Transformers · Anthropic

Notes

• AI Prompt Engineering Is Dead — Automated prompt optimization outperforms human prompt engineering
• AMD Researchers Introduce Agent Laboratory: An Autonomous LLM-based Framework Capable of Completing the Entire Research Process — Autonomous LLM pipeline completing literature review, experiments, and paper writing
• About — AAAI 2024 workshop on cooperative multi-agent decision-making and learning
• An Introduction to Building Custom Reinforcement Learning Environments Using OpenAI Gym — Tutorial for building custom OpenAI Gym RL environment
• An Opinionated Guide to ML Research — Practical advice on problem selection and research habits for ML researchers
• CAMEL-AI | Finding the Scaling Laws of Agents — CAMEL-AI open-source multi-agent framework for research on agent scaling laws
• CS285 — UC Berkeley graduate deep reinforcement learning course lecture index
• Can 1B LLM Surpass 405B LLM? Optimizing Computation for Small LLMs to Outperform Larger Models — Test-time scaling lets small LLMs outperform much larger models
• ChatGPT is 'not particularly innovative,' and 'nothing revolutionary', says Meta's chief AI scientist — Yann LeCun argues ChatGPT is solid engineering not scientific breakthrough
• Decisions from Data: How Offline Reinforcement Learning Will Change How We Use ML — Argument for offline RL enabling data-driven end-to-end decision making
• Deep Reinforcement Learning Doesn't Work Yet — Systematic critique of deep RL limitations and failure modes
• Deep Reinforcement Learning at the Edge of the Statistical Precipice — Critique of deep RL evaluation using point estimates; proposes robust statistical methodology
• DeepMind scientists: Reinforcement learning is enough for general AI — DeepMind paper arguing reward maximization alone yields AGI
• DeepMind's Collect & Infer: A Fresh Look at Data-Efficient Reinforcement Learning | Synced — DeepMind Collect and Infer paradigm separates RL data collection from inference
• Flappy Bird RL — Q-learning agent trained to play Flappy Bird in JavaScript
• From Data to Drugs: The Role of Artificial Intelligence in Drug Discovery — AI and foundation models transforming pharmaceutical drug discovery pipelines
• GitHub - RL-Research-Cohiba/Reinforcement_Learning: Reinforcement Learning research — GitHub repo collecting reinforcement learning research resources and links
• GitHub - catalyst-team/dl-course: Deep Learning with Catalyst — Open-source deep learning course using PyTorch and Catalyst framework
• GitHub - pfnet/pfrl: PFRL: a PyTorch-based deep reinforcement learning library — PyTorch open-source library implementing state-of-the-art deep RL algorithms
• Goal-directed graph construction using reinforcement learning — RL agent with GNNs constructs graphs optimizing target structural properties
• Google AI Introduces ArchGym: An Open-Source Gymnasium for Machine Learning that Connects a Diverse Range of Search Algorithms To Architecture Simulators — Open-source ML gymnasium for reproducible computer architecture design space exploration
• How the Pandemic Made Algorithms Go Haywire — COVID-19 distribution shift caused healthcare and finance ML algorithms to fail
• Hugging Face Introduces StackLLaMA: A 7B Parameter Language Model Based on LLaMA and Trained on Data from Stack Exchange Using RLHF — Hugging Face RLHF fine-tuning of LLaMA 7B on Stack Exchange Q&A data
• Image Augmentation Is All You Need: Regularizing Deep Reinforcement Learning from Pixels — Data augmentation technique enabling model-free RL directly from pixels
• Intuitive RL: Intro to Advantage-Actor-Critic (A2C) | HackerNoon — Intuitive narrative introduction to the Advantage-Actor-Critic RL algorithm
• Machine Learning Could Create the Perfect Game Bosses — ML-driven NPC behavior in games balances difficulty tuning over scripted trees
• Machine Learning Course Series — Three-course ML curriculum from Python basics to LLMs
• Machine Learning Course Series — Three-course ML curriculum from Python basics to LLMs
• Multi-Armed Bandits in Python: Epsilon Greedy, UCB1, Bayesian UCB, and EXP3 — Four bandit algorithm implementations evaluated on MovieLens recommendation task
• Neglected machine learning ideas — Survey of underrepresented machine learning techniques missing from textbooks
• NeurIPS 2023: Key Takeaways From Invited Talks — NeurIPS 2023 invited talk summaries on LLM efficiency, generative AI, and responsible data
• New Go-playing trick defeats world-class Go AI—but loses to human amateurs — Adversarial policy exploits off-distribution moves to beat world-class Go AI
• Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problems — Survey of offline RL algorithms learning policies from static datasets without online interaction
• Optimism in the Face of Uncertainty: the UCB1 Algorithm — Mathematical derivation and implementation of the UCB1 bandit algorithm
• REINFORCEMENT LEARNING AND OPTIMAL CONTROL — Bertsekas textbook unifying RL, dynamic programming, and optimal control
• RL4HCI — Workshop agenda building RL research agenda for human-computer interaction
• Rainbow: Combining Improvements in Deep Reinforcement Learning — Empirical combination of six DQN extensions achieving state-of-the-art Atari performance
• Real World Applications of Markov Decision Process (MDP) — Real-world finite MDP formulations with worked transition graph examples
• Reinforcement Learning algorithms — an intuitive overview — Survey of model-free and model-based RL algorithm families
• Researchers From Stanford And DeepMind Come Up With The Idea of Using Large Language Models LLMs as a Proxy Reward Function — LLMs as proxy reward functions for RL agent alignment via natural language
• Scaling Recommender Transformers to a Billion Parameters | Towards Data Science — ARGUS autoregressive recommender transformer scaled to one billion parameters
• Self-supervised learning is the future of AI — LeCun's blueprint for self-supervised learning over supervised AI
• Self-supervised learning: The plan to make deep learning data-efficient - TechTalks — Yann LeCun's AAAI 2020 roadmap for data-efficient self-supervised learning
• Silicon Valley bets big on 'environments' to train AI agents | TechCrunch — RL environments emerge as critical training infrastructure for capable AI agents
• Study: Machine learning a useful tool for quantum control — Reinforcement learning enables real-time quantum feedback control under noise
• The Era of Agentic Organization: Learning to Organize with Language Models — AsyncThink paradigm: concurrent LLM reasoning optimized via reinforcement learning
• The challenges of reinforcement learning from human feedback (RLHF) - TechTalks — RLHF limitations across feedback, reward modeling, and policy
• The second wave of AI coding is here — Next-gen AI coding agents targeting functional correctness via process data
• Towards Data Science — Benchmark on latency, cost, reproducibility for AI agents
• We Got Claude to Fine-Tune an Open Source LLM — Hugging Face skill enables Claude Code to submit and manage LLM training jobs
• What AI Models for War Actually Look Like — Military-specialized AI startup building models for mission planning and decision dominance
• What is Reinforcement Learning and 9 examples of what you can do with it. — Introductory overview of reinforcement learning with real-world examples
• Why machine learning struggles with causality - TechTalks — ML systems lack causal reasoning needed for robust generalization