Study shows mutual information can reward structure learning agents without expert systems.
problem Designing rewards for structure learning agents in natural language environments.
method Revisited Information Theory of unsupervised induction of phrase-structure grammars, using random sets of linguistic samples.
result Empirical evidence that simulated semantic structures can be distinguished from random ones by mutual information among their constituents.
Paper introduces PRMs to learn non-Markovian stochastic rewards for reinforcement learning.
problem Lack of structured representation for non-Markovian stochastic rewards in reinforcement learning.
method Introduces probabilistic reward machines (PRMs) and presents an algorithm to learn them from decision processes.
result Algorithm proves correct and convergent for learning PRMs from decision processes.
A model learns rewards from diverse demonstrations for structurally similar tasks.
problem Difficulty in learning reward functions from demonstrations in real-world applications.
method Deep latent variable model that infers rewards from heterogeneous demonstrations of related tasks.
result Model can infer rewards for new tasks from a single demonstration.
PixL2R maps natural language to pixel-based rewards for RL, improving sample efficiency.
problem Sparse reward settings in RL limit applicability to complex problems.
method Directly maps natural language descriptions to pixel-based rewards for guiding RL.
result Language-based rewards significantly improve sample efficiency in policy learning.
Proposes a method to boost deep reinforcement learning with sparse rewards.
problem Challenges in learning complex behaviors with long horizons and sparse rewards.
method Predictive coding for reward shaping.
result Achieves better learning by providing reward signals that understand environment dynamics and emphasize useful features.
New IRL algorithm identifies optimal reward and policy from expert demonstrations.
problem Understanding reward functions from expert demonstrations with neural networks.
method Two-timescale single-loop IRL algorithm for neural network parameterized rewards.
result First IRL algorithm with non-asymptotic convergence guarantee and global optimality in neural network settings.
Study reward-free RL in non-linear settings, improving efficiency and removing assumptions.
problem Improving sample efficiency in reward-free reinforcement learning for non-linear function approximation.
method Proposed RFOLIVE algorithm for minimal structural assumptions, analyzed hardness results for reward-free and reward-aware exploration.
result Statistical efficiency and hardness results under various structural assumptions, no need for reachability or explorability assumptions.
A new IRL model recovers reward and state structure from expert demonstrations.
problem Limitation of classical maximum entropy model in capturing state structure.
method Generalized maximum causal entropy for IRL models.
result Empirically outperforms classical models in recovering reward and state structure.
Reinforcement Learning (RL) agents require the specification of a reward signal for learning behaviours. However, introduction of corrupt or stochastic rewards can yield high variance in learning. Such corruption may be a direct result of goal misspecification, randomness in the reward signal, or correlation of the rew…
Paper simplifies complex AI exploration by predicting future rewards.
problem Training machines to optimally gather complex information.
method Developed a denser reward structure using cross-value to decouple exploration and exploitation.
result Demonstrated successful learning of challenging tasks without shaping or bonuses.
Enhances reward specification in RL with a novel language-based approach.
problem Reward specification in RL can lead to unintended, potentially harmful behaviours.
method Developed a novel class of language-based Reward Machines using RML's built-in memory.
result Can specify non-regular, non-Markovian reward functions for complex tasks.
New bandit algorithms improve sparse reward learning.
problem Sparse rewards hinder learning efficiency in real-world bandit applications.
method Developed algorithms based on Upper Confidence Bound and Thompson Sampling for zero-inflated distributions.
result Empirical performance of new algorithms is superior to existing methods.
We consider a problem of learning the reward and policy from expert examples under unknown dynamics. Our proposed method builds on the framework of generative adversarial networks and introduces the empowerment-regularized maximum-entropy inverse reinforcement learning to learn near-optimal rewards and policies. Empowe…
A new algorithm tackles delayed combinatorial semi-bandit with causal relations.
problem Optimizing decisions in a non-stationary environment with delayed and causally related rewards.
method Formalized as a non-stationary delayed combinatorial semi-bandit problem, the approach models causal relations with a directed graph in a stationary structural equation model. The agent learns these relations from delayed feedback to optimize decisions.
result Proved a regret bound for the proposed algorithm's performance.
New algorithm uses Whittle index to improve Q-learning for restless bandits.
problem Optimizing decision-making in multiarmed restless bandits with average reward.
method Introduces a novel reinforcement learning algorithm combining Q-learning and Whittle index policy.
result Demonstrates significant computational gains and excellent empirical performance.
In structured output prediction tasks, labeling ground-truth training output is often expensive. However, for many tasks, even when the true output is unknown, we can evaluate predictions using a scalar reward function, which may be easily assembled from human knowledge or non-differentiable pipelines. But searching th…
The paper tackles reward-relevance in offline RL with sparse decision dynamics.
problem Offline reinforcement learning with sparse decision dynamics and estimation sparsity.
method Reward-filtered least-squares policy evaluation using thresholded lasso.
result The method provides theoretical guarantees with sample complexity dependent on sparse component size.
Trade-R1 bridges verifiable rewards to stochastic financial markets via process-level reasoning verification.
problem Extending RL to financial markets where rewards are verifiable but noisy.
method A verification method that transforms reasoning over financial documents into a structured RAG task, using a triangular consistency metric.
result DSR achieves superior cross-market generalization while maintaining reasoning consistency.
Extends linear MDP to handle nonlinear rewards.
problem Restrictive linear MDP assumption limits real-world applicability.
method Proposes Generalized Linear MDP (GLMDP) with GLMs for rewards.
result Develops offline RL algorithms achieving suboptimality guarantees.
We consider an adversarial online learning setting where a decision maker can choose an action in every stage of the game. In addition to observing the reward of the chosen action, the decision maker gets side observations on the reward he would have obtained had he chosen some of the other actions. The observation str…
Study shows fast rates for inverse reinforcement learning with linear rewards.
problem Entropy-regularized min-max inverse reinforcement learning in finite-horizon MDPs.
method Structural and statistical analysis of Min-Max-IRL with pseudo-self-concordance.
result Both trajectory-level KL divergence and parameter error decay at O ( n − 1 ) \mathcal{O}(n^{-1}) O ( n − 1 ) . Adversarial RL recovers agent rewards from financial market data simulations.
problem Recovering agent rewards in volatile financial markets with unknown dynamics.
method Adversarial inverse reinforcement learning in latent space simulations.
result Adversarial RL can robustly recover agent rewards from latent space representations of real market data.
Reinforcement learning is concerned with identifying reward-maximizing behaviour policies in environments that are initially unknown. State-of-the-art reinforcement learning approaches, such as deep Q-networks, are model-free and learn to act effectively across a wide range of environments such as Atari games, but requ…
Reinforcement learning is widely used for dialogue policy optimization where the reward function often consists of more than one component, e.g., the dialogue success and the dialogue length. In this work, we propose a structured method for finding a good balance between these components by searching for the optimal re…
Novel algorithm reduces computational burden in IRL with finite-time guarantees.
problem Efficiently recover reward function and optimal policy from expert behavior.
method Single-loop algorithm that maximizes likelihood after each policy improvement step.
result Algorithm provably converges to a stationary solution with finite-time guarantees.
New algorithm DUSA minimizes regret in structured bandits with structural information.
problem Optimal decision-making under uncertainty with structural reward information.
method DUSA algorithm exploiting convex structural information.
result Regret matches information-theoretic lower bound up to a constant factor.
DeepSynth synthesizes automata to guide deep RL agents through sparse, non-Markovian rewards.
problem Training deep RL agents with sparse, non-Markovian rewards and unknown high-level objectives.
method Employing a novel algorithm for synthesizing compact automata to uncover sequential structure from trace data.
result Reduces the number of iterations required for policy synthesis by two orders of magnitude and improves scalability.
Two algorithms tackle heavy-tailed rewards in reinforcement learning with linear function approximation.
problem Online sequential decision-making with heavy-tailed rewards.
method AdaOFUL and VARA algorithms for linear stochastic bandits and MDPs, using modified adaptive Huber regression.
result Achieved state-of-the-art and variance-aware regret bounds for heavy-tailed rewards.
Consider a nonparametric contextual multi-arm bandit problem where each arm a ∈ [ K ] a \in [K] a ∈ [ K ] is associated to a nonparametric reward function f a : [ 0 , 1 ] → R f_a: [0,1] \to \mathbb{R} f a : [ 0 , 1 ] → R mapping from contexts to the expected reward. Suppose that there is a large set of arms, yet there is a simple but unknown structure amongst the arm reward…
New algorithm reduces regret in multi-armed bandit problems with Gaussian rewards.
problem Optimizing decisions in multi-armed bandit problems with Gaussian rewards.
method Proposed TSCG and UTSCG algorithms using Thompson Sampling with Gaussian prior.
result Achieved lower regret bounds for optimal arm selection.
New algorithm reduces reinforcement learning regret to sqrt(T) without strong dynamics assumptions.
problem Infinite-horizon average-reward reinforcement learning with linear MDPs.
method Approximate by discounted-reward MDPs and apply optimistic value iteration.
result Achieves O(sqrt(T)) regret with polynomial complexity.
A new estimator for state values in reinforcement learning reduces complexity and improves convergence.
problem Estimating state values in reinforcement learning with Markov reward processes.
method Loop estimator exploiting regenerative structure of Markov reward processes.
result Instance-dependent convergence rate of O ~ ( τ s / T ) \widetilde{O}\left(\sqrt{τ_s/T}\right) O ( τ s / T ) for estimating state values. The explore{exploit dilemma is one of the central challenges in Reinforcement Learning (RL). Bayesian RL solves the dilemma by providing the agent with information in the form of a prior distribution over environments; however, full Bayesian planning is intractable. Planning with the mean MDP is a common myopic approxi…
Paper addresses reward learning issues in RL, improving both under- and over-estimation.
problem Reward learning from data can lead to reward delusions or underestimation, causing unintended behaviors.
method Connects reward learning to positive-unlabeled (PU) learning and applies a large-scale PU learning algorithm.
result Improves both GAIL and supervised reward learning without additional assumptions.
This paper explores how environmental properties can simplify reinforcement learning in non-episodic settings.
problem Challenges in reinforcement learning with continuous interaction and sparse delayed rewards.
method Analysis of environment shaping and dynamism properties to simplify learning.
result Properties like environment shaping and dynamism can significantly ease learning in non-episodic, sparse reward settings.
SA-GFN corrects biases in GFlowNets due to graph symmetries.
problem Systematic biases in state transition probability computations.
method Incorporates symmetry corrections into the learning process through reward scaling.
result Eliminates need for explicit state transition computations.
Algorithm estimates human decision-making in high-dimensional states with finite-time guarantees.
problem Estimating optimal policies and measures of fit in dynamic decision models with high-dimensional state spaces.
method Single-loop estimation algorithm with stochastic gradient steps for reward maximization.
result Algorithm converges to a stationary solution with finite-time guarantees and approximates maximum likelihood sublinearly.
A scalable MARL algorithm using local rewards for cooperative multi-agent learning.
problem Scalability issues in cooperative multi-agent reinforcement learning due to large state and action spaces.
method LOMAQ algorithm incorporating local rewards in centralized training and decentralized execution.
result LOMAQ scales well compared to other methods, improving performance and convergence speed.
Privacy-preserving reinforcement learning from human feedback using decoupled reward modeling.
problem Training large language models with sensitive user information while preserving privacy.
method Proposes a privacy-preserving framework that imposes differential privacy on reward learning only.
result Privacy contributes an additional additive term to the suboptimality gap, and the upper bound is rate-optimal up to logarithmic factors.
Self-supervised reward prediction improves RL in sparse reward settings.
problem Data efficiency and sparse reward signals in reinforcement learning.
method Learning a state representation for reward prediction and using it to shape rewards.
result Self-supervised reward prediction enhances RL algorithms in single-goal environments.
Paper proposes RRD to learn proxy rewards for sparse delayed rewards in episodic reinforcement learning.
problem Learning from sparse and delayed rewards in reinforcement learning.
method Randomized Return Decomposition (RRD) algorithm to redistribute rewards.
result Substantial improvement over baseline algorithms in experiments.
New method optimizes policies without assuming known link functions between preferences and rewards.
problem Policy alignment with unknown and unrestricted link functions.
method Formulates an f f f -divergence-constrained reward maximization problem, learning policies directly. result Induces a semiparametric single-index binary choice model for policy alignment.
A reinforcement learning framework for Mars rover control using temporal logic.
problem Sparse rewards in continuous-state continuous-action MDPs with high-level temporal structures.
method Actor-critic, model-free, online RL framework with modular DDPG architecture.
result Success rate of synthesised policy in Mars rover experiment.
This paper improves understanding of GAIL's generalization and computational efficiency.
problem Understanding the theoretical properties of GAIL, especially its generalization and computational aspects.
method Investigates GAIL's theoretical properties, showing guarantees for generalization and computational efficiency.
result GAIL can be efficiently solved by stochastic first order optimization algorithms with sublinear convergence.
Flexible algorithms for maximizing rewards in structured bandits.
problem Reward maximization in structured stochastic multi-armed bandit problems.
method Asymptotically optimal algorithms using iterative saddle-point solvers.
result Achieves optimal performance with minimal computational burden.
This paper introduces a new reward shaping method for average-reward reinforcement learning.
problem Speeding up convergence to an optimal policy in average-reward reinforcement learning tasks.
method Developed a temporal logic-based approach to automatically generate reward shaping functions.
result The optimal policy can be recovered using the proposed reward shaping framework.
Generative Flow Networks use submodular upper bounds to generate more data.
problem Generating data from unknown, complex reward functions efficiently.
method Introduce submodular upper bounds to estimate reward, use Optimism in the Face of Uncertainty principle to train GFNs.
result SUBo-GFN generates significantly more data than classical GFNs.
SS-SARSA tackles recovering bandits by treating rounds as states.
problem Real-world rewards vary with the elapsed rounds since last pull.
method Proposes SS-SARSA, a new RL algorithm treating rounds as states.
result Achieves efficient learning with asymptotic convergence to optimal policy.