Best AI papers explained

441 Episodes

1. How Bidirectionality Helps Language Models Learn Better via Dynamic Bottleneck Estimation

   Published: 6/6/2025
2. A Closer Look at Bias and Chain-of-Thought Faithfulness of Large (Vision) Language Models

   Published: 6/5/2025
3. Simplifying Bayesian Optimization Via In-Context Direct Optimum Sampling

   Published: 6/5/2025
4. Bayesian Teaching Enables Probabilistic Reasoning in Large Language Models

   Published: 6/5/2025
5. IPO: Interpretable Prompt Optimization for Vision-Language Models

   Published: 6/5/2025
6. Evolutionary Prompt Optimization discovers emergent multimodal reasoning strategies

   Published: 6/5/2025
7. Evaluating the Unseen Capabilities: How Many Theorems Do LLMs Know?

   Published: 6/4/2025
8. Diffusion Guidance Is a Controllable Policy Improvement Operator

   Published: 6/2/2025
9. Alita: Generalist Agent With Self-Evolution

   Published: 6/2/2025
10. A Snapshot of Influence: A Local Data Attribution Framework for Online Reinforcement Learning

    Published: 6/2/2025
11. Learning Compositional Functions with Transformers from Easy-to-Hard Data

    Published: 6/2/2025
12. Preference Learning with Response Time

    Published: 6/2/2025
13. Accelerating RL for LLM Reasoning with Optimal Advantage Regression

    Published: 5/31/2025
14. Algorithms for reliable decision-making need causal reasoning

    Published: 5/31/2025
15. Belief Attribution as Mental Explanation: The Role of Accuracy, Informativity, and Causality

    Published: 5/31/2025
16. Distances for Markov chains from sample streams

    Published: 5/31/2025
17. When and Why LLMs Fail to Reason Globally

    Published: 5/31/2025
18. IDA-Bench: Evaluating LLMs on Interactive Guided Data Analysis

    Published: 5/31/2025
19. No Free Lunch: Non-Asymptotic Analysis of Prediction-Powered Inference

    Published: 5/31/2025
20. Accelerating RL for LLM Reasoning with Optimal Advantage Regression

    Published: 5/31/2025