Get Started
Home
Topics
Search
Library
Research questionHow can we learn near-optimal policies from transition samples in large or infinite-state-action MDPs using function approximation?Tabular policy representations become infeasible when an MDP has a large or infinite state-action space. The challenge is to use function approximation to learn reliable policies from transition samples without requiring an impractical number of queries.
AI
Economics
Finance
Machine Learning
Reinforcement Learning
Research Paper
Statistical Machine Learning
Latest papersRecent research connected to this question, newest first.Adaptive Partitioning and Learning for Stochastic Control of Diffusion ProcessesThe study considers bounded continuous actions, polynomially growing rewards, and unknown drift, volatility, and reward functions in controlled diffusion processes. It presents theoretical regret guarantees for an adaptive state-action partitioning approach and reports numerical experiments, including high-dimensional multi-asset mean-variance portfolio selection.research paper · Sep 3, 2026Adaptive Resolving Methods for Markov Decision Processes with Function ApproximationsApplies to online and data-driven MDP learning with function approximation. The evidence covers an LP-based algorithm, repeated resolving as transition samples arrive, instance-dependent and robust convergence guarantees, separate accounting of historical samples and resolving-round queries, and numerical experiments; it does not specify a particular function class or deployment environment.research paper · Sep 3, 2026
Related questions
How can reinforcement-learning data acquisition be optimized when interactions are costly, slow, and human-mediated?How can policy optimization for long-horizon LLM agents preserve useful transitions across updates when rollout groups are small?How can offline reinforcement learning improve policies beyond dataset support while keeping value estimates reliable under distribution shift?How can we learn welfare-optimal approximate Nash equilibria in concurrent stochastic games under uncertain transitions while certifying exact-equilibrium nonexistence?