On the Computational Tractability of Robust Bandits
cs.LG
Submitted: 2026-10-06
Updated: 2026-10-06
Terminology
Sources
- Concrete Problems in AI Safety
- Occam's razor is insufficient to infer the preferences of irrational agents
- Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback
- Robust Cooperation in the Prisoner's Dilemma: Program Equilibrium via Provability Logic
- Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback
- Deep reinforcement learning from human preferences
- Embedded Agency
- Reflective Oracles: A Foundation for Classical Game Theory
- The Statistical Complexity of Interactive Decision Making
- Cooperative Inverse Reinforcement Learning
- Inverse Reward Design
- Risks from Learned Optimization in Advanced Machine Learning Systems
- Reward-rational (implicit) choice: A unifying formalism for reward learning
- When is Agnostic Reinforcement Learning Statistically Tractable?
- A Formal Solution to the Grain of Truth Problem
- Scalable agent alignment via reward modeling: a research direction
- The Alignment Problem from a Deep Learning Perspective
- Learning with Multiple Correct Answers -- Regret Bounds under Different Feedback Models
- Preferences Implicit in the State of the World
- Limit-Computable Grains of Truth for Arbitrary Computable Extensive-Form (Un)Known Games
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks