Explore
Tags
Browse articles across categories by shared concepts, technologies, formats, and other controlled classifications.
Sequential Decision-Making
Core reinforcement-learning framing in which actions affect future states and outcomes, including distinctions from one-step prediction or decision settings.
Browse articlesStates and Observations
State representations, observations, Markov sufficiency, hidden information, and the information available to an agent when choosing actions.
Browse articlesRewards and Returns
Reward signals, accumulated return objectives, discounting, and the relationship between immediate feedback and long-run outcomes.
Browse articlesPolicies and Control
Policies as decision rules and the process of evaluating, improving, or optimizing behavior.
Browse articlesValue Functions
State values, action values, advantages, and related predictions of future outcomes used to compare decisions.
Browse articlesBellman Reasoning
Bellman equations, backups, fixed points, and recursive value reasoning used to connect immediate outcomes with continuation value.
Browse articlesExploration
How agents seek information or discover useful behavior, including exploration strategies, bonuses, uncertainty, and intrinsic motivation.
Browse articlesTemporal-Difference Learning
Learning from sampled experience using bootstrapped estimates, including TD targets, multi-step returns, and eligibility traces.
Browse articlesModel-Based Reinforcement Learning
Learning or using transition and reward models for planning, including model error and the tradeoff between imagined and real experience.
Browse articlesPartial Observability
Decision-making when observations do not reveal the full state, including beliefs, memory, and POMDP reasoning.
Browse articlesPolicy Gradients
Direct policy optimization using gradients of expected outcomes, including estimator construction, baselines, and actor-critic updates.
Browse articlesFunction Approximation
Using parameterized representations such as neural networks instead of tables, including generalization and approximation-related behavior.
Browse articlesExperimental Evaluation
Designing, measuring, comparing, and diagnosing RL experiments with repeatable procedures, suitable metrics, and uncertainty-aware interpretation.
Browse articlesReward Design
Designing, shaping, learning, or auditing reward signals and proxies, including the risks of objectives that diverge from intended outcomes.
Browse articlesOffline Reinforcement Learning
Learning or evaluating policies from fixed logged experience without relying on unrestricted new interaction, including support and extrapolation limits.
Browse articlesSafety and Constraints
Representing, enforcing, or evaluating safety, resource, and risk requirements in reinforcement-learning behavior.
Browse articlesMulti-Agent Learning
Learning and decision-making involving multiple interacting agents, including coordination, competition, non-stationarity, communication, and agent-level credit.
Browse articlesImitation and Demonstrations
Learning from expert behavior or demonstrations, including behavior cloning, covariate shift, inverse reward inference, and combining demonstrations with environment feedback.
Browse articlesPreference-Based Learning
Using comparative judgments or preferences to infer reward-like objectives and guide reinforcement-learning optimization.
Browse articlesContinuous Control
Reinforcement learning with continuous action spaces, including action distributions, bounded actions, and continuous-control actor-critic methods.
Browse articlesPractice
Runnable implementations and experiments intended for readers to execute, inspect, or modify as a hands-on learning activity.
Browse articles