arXiv research

Study optimal and equitable encouragement policies for treatment adherence.

Paper introduces EnCounteR for estimating causal effects using encouragement data.

Standard reinforcement learning methods aim to master one way of solving a task whereas there may exist multiple near-optimal policies. Being able to identify this collection of near-optimal policies can allow a domain expert to efficiently explore the space of reasonable solutions. Unfortunately, existing approaches t…

Softmax policy gradient achieves global optimality in wide neural networks with entropy regularization.

Entropy regularization is commonly used to improve policy optimization in reinforcement learning. It is believed to help with \emph{exploration} by encouraging the selection of more stochastic policies. In this work, we analyze this claim using new visualizations of the optimization landscape based on randomly perturbi…

Policy optimization is a core component of reinforcement learning (RL), and most existing RL methods directly optimize parameters of a policy based on maximizing the expected total reward, or its surrogate. Though often achieving encouraging empirical success, its underlying mathematical principle on {\em policy-distri…

A new algorithm trains experts to safely guide agents in partially observed environments.

This work improves policy-based training by proposing an evaluation balance objective for GFlowNets.

Entropy-regularized NPG methods converge linearly in discounted MDPs.

This work explains why online imitation learning improves faster than theory predicts.

Curriculum learning and imitation learning improve control over financial time-series data.

Optimistic NPG improves policy optimization in online RL with efficient sample complexity.

This work improves policy optimization by maximizing entropy of state distribution, leading to better exploration.

Study finds dividend payout policy positively impacts firm profitability.

A single policy suffices for near-optimal parallel exploration in RL.

New method boosts skill learning by encouraging optimistic exploration.

This paper explores a simple regularizer for reinforcement learning by proposing Generative Adversarial Self-Imitation Learning (GASIL), which encourages the agent to imitate past good trajectories via generative adversarial imitation learning framework. Instead of directly maximizing rewards, GASIL focuses on reproduc…

Text generation is a crucial task in NLP. Recently, several adversarial generative models have been proposed to improve the exposure bias problem in text generation. Though these models gain great success, they still suffer from the problems of reward sparsity and mode collapse. In order to address these two problems, …

POMBU improves model-based RL's asymptotic performance by estimating and using uncertainty.

ConQUR tackles delusional bias in deep Q-learning, improving performance in Atari games.

GPMD solves regularized RL with linear convergence, promoting structural policies.

Paper connects DP and optimization for RL, suggesting new algorithms.

Self-consistent models improve reinforcement learning by aligning predictions with future values.

Cyclical learning rate improves neural machine translation performance.

Regularization improves policy optimization in RL, especially on harder tasks.

Encouraging sustainable mobility patterns is at the forefront of policymaking at all scales of governance as the collective consciousness surrounding climate change continues to expand. Not every community, however, possesses the necessary economic or socio-cultural capital to encourage modal shifts away from private m…

Maximizes Rényi entropy for efficient exploration in reward-free RL.

Improved RL policies from offline data with relaxed BC constraints.

Sparse reward is one of the biggest challenges in reinforcement learning (RL). In this paper, we propose a novel method called Generative Exploration and Exploitation (GENE) to overcome sparse reward. GENE automatically generates start states to encourage the agent to explore the environment and to exploit received rew…

Two DRL policies collaborate to solve NP-hard routing problems.

Reinforcement learning with sparse rewards is challenging because an agent can rarely obtain non-zero rewards and hence, gradient-based optimization of parameterized policies can be incremental and slow. Recent work demonstrated that using a memory buffer of previous successful trajectories can result in more effective…

In Multi-Goal Reinforcement Learning, an agent learns to achieve multiple goals with a goal-conditioned policy. During learning, the agent first collects the trajectories into a replay buffer, and later these trajectories are selected randomly for replay. However, the achieved goals in the replay buffer are often biase…

Improves BC policies by generating new plausible trajectories.

In recent years, deep reinforcement learning has been shown to be adept at solving sequential decision processes with high-dimensional state spaces such as in the Atari games. Many reinforcement learning problems, however, involve high-dimensional discrete action spaces as well as high-dimensional state spaces. This pa…

Self-guided ALPs improve MDP policies without domain knowledge.

A framework for navigating environments with spatially correlated obstacles and uncertain blockage status.

UCPO improves diversity in reinforcement learning models, maintaining high accuracy.

Learning to cooperate with friends and compete with foes is a key component of multi-agent reinforcement learning. Typically to do so, one requires access to either a model of or interaction with the other agent(s). Here we show how to learn effective strategies for cooperation and competition in an asymmetric informat…

Empowerment is an information-theoretic method that can be used to intrinsically motivate learning agents. It attempts to maximize an agent's control over the environment by encouraging visiting states with a large number of reachable next states. Empowered learning has been shown to lead to complex behaviors, without …

LICA learns credit assignment for cooperative agents without explicit formulation.

Agent learns directed exploration policies to improve performance in hard games.

Study multi-task learning with low-rank representation in stochastic linear bandits.

Novel RL-based NPG improves multi-objective NAS efficiency and performance.

The task of dialog management is commonly decomposed into two sequential subtasks: dialog state tracking and dialog policy learning. In an end-to-end dialog system, the aim of dialog state tracking is to accurately estimate the true dialog state from noisy observations produced by the speech recognition and the natural…

Partially observable environments present an important open challenge in the domain of sequential control learning with delayed rewards. Despite numerous attempts during the two last decades, the majority of reinforcement learning algorithms and associated approximate models, applied to this context, still assume Marko…

ReMixMatch improves semi-supervised learning with new techniques for data efficiency.

Model-free reinforcement learning methods such as the Proximal Policy Optimization algorithm (PPO) have successfully applied in complex decision-making problems such as Atari games. However, these methods suffer from high variances and high sample complexity. On the other hand, model-based reinforcement learning method…

Mutual-information regularization improves RL algorithms, especially in high-dimensional domains.