Training AI Without Writing A Reward Function, with Reward Modelling Published 2019-12-13 Download video MP4 360p Recommendations 10:22 AI That Doesn't Try Too Hard - Maximizers and Satisficers 23:24 The OTHER AI Alignment Problem: Mesa-Optimizers and Inner Alignment 16:00 AI Safety Gym - Computerphile 16:29 10 Reasons to Ignore AI Safety 11:47 We Were Right! Real Inner Misalignment 25:57 GPT3: An Even Bigger Language Model - Computerphile 11:32 How to Keep Improving When You're Better Than Any Teacher - Iterated Distillation and Amplification 16:56 Vectoring Words (Word Embeddings) - Computerphile 09:24 Why Does AI Lie, and What Can We Do About It? 15:38 A Response to Steven Pinker on AI 09:40 9 Examples of Specification Gaming 20:41 Training an unbeatable AI in Trackmania 23:45 Stop Button Solution? - Computerphile 26:09 Computer Scientist Explains Machine Learning in 5 Levels of Difficulty | WIRED 09:54 Quantilizers: AI That Doesn't Try Too Hard 20:18 Why Does Diffusion Work Better than Auto-Regression? 10:20 Deceptive Misaligned Mesa-Optimisers? It's More Likely Than You Think... 59:17 RLHF: How to Learn from Human Feedback with Reinforcement Learning 45:59 AI Ruined My Year 10:41 What can AGI do? I/O and Speed Similar videos 08:40 AI Learns to Walk (deep reinforcement learning) 00:20 Writing successful reward functions 08:55 Direct Preference Optimization: Your Language Model is Secretly a Reward Model | DPO paper explained 00:19 Q Learning Rewards Modeling 00:46 NEVER buy from the Dark Web.. #shorts 07:54 How ChatGPT Works Technically | ChatGPT Architecture 12:16 Does your PPO agent fail to learn? 15:55 Reinforcement Learning Made Simple - Reward 09:10 Direct Preference Optimization: Forget RLHF (PPO) 05:58 Reward Hacking in AI 1:00:38 Reinforcement Learning from Human Feedback: From Zero to chatGPT 33:53 Training AI to Play Pokemon with Reinforcement Learning 03:23 How to Design a Reinforcement Learning Reward Function for a Lunar Lander 🛸 08:55 How AIs, like ChatGPT, Learn 03:12 Reward Model for RLHF with Google Colab + trl 1:16:18 Stanford ECON295/CS323 I 2024 I AI and Creativity, Anima Anandkumar 18:49 Text2Reward: Automated Dense Reward Function Generation for Reinforcement Learning 11:59 How ChatGPT is Trained - model and training explained 52:03 Edward Grefenstette: Teaching Artificial Agents to Understand Language by Modelling Reward More results