60 citations · 153 across the 26 of their papers we have counts for
14 papers · 1 filter
Random Policy Enables In-Context Reinforcement Learning within Trust Horizons
Weiqin Chen, Santiago Paternain
Pretrained foundation models have exhibited extraordinary in-context learning performance, allowing zero-shot generalization to new tasks not encountered during pretraining. In the…
Provable Domain Adaptation for Offline Reinforcement Learning with Limited Samples
Weiqin Chen, Xinjie Zhang, Sandipan Mishra +1
Offline reinforcement learning (RL) learns effective policies from a static target dataset. The performance of state-of-the-art offline RL algorithms notwithstanding, it relies on…
A General Control-Theoretic Approach for Reinforcement Learning: Theory and Algorithms
Weiqin Chen, Mark S. Squillante, Chai Wah Wu +1
We devise a control-theoretic reinforcement learning approach to support direct learning of the optimal policy. We establish various theoretical properties of our approach, such as…
Adaptive Primal-Dual Method for Safe Reinforcement Learning
Weiqin Chen, James Onyejizu, Long Vu +5
Primal-dual methods have a natural application in Safe Reinforcement Learning (SRL), posed as a constrained policy optimization problem. In practice however, applying primal-dual m…
Probabilistic Constraint for Safety-Critical Reinforcement Learning
Weiqin Chen, Dharmashankar Subramanian, Santiago Paternain
In this paper, we consider the problem of learning safe policies for probabilistic-constrained reinforcement learning (RL). Specifically, a safe policy or controller is one that, w…
Policy Gradients for Probabilistic Constrained Reinforcement Learning
Weiqin Chen, Dharmashankar Subramanian, Santiago Paternain
This paper considers the problem of learning safe policies in the context of reinforcement learning (RL). In particular, we consider the notion of probabilistic safety. This is, we…
Constrained Learning with Non-Convex Losses
Luiz F. O. Chamon, Santiago Paternain, Miguel Calvo-Fullana +1
Though learning has become a core component of modern information processing, there is now ample evidence that it can lead to biased, unsafe, and prejudiced systems. The need to im…
Towards Safe Continuing Task Reinforcement Learning
Miguel Calvo-Fullana, Luiz F. O. Chamon, Santiago Paternain
Safety is a critical feature of controller design for physical systems. When designing control policies, several approaches to guarantee this aspect of autonomy have been proposed,…
State Augmented Constrained Reinforcement Learning: Overcoming the Limitations of Learning with Rewards
Miguel Calvo-Fullana, Santiago Paternain, Luiz F. O. Chamon +1
A common formulation of constrained reinforcement learning involves multiple rewards that must individually accumulate to given thresholds. In this class of problems, we show a sim…
Trust but Verify: Assigning Prediction Credibility by Counterfactual Constrained Learning
Luiz F. O. Chamon, Santiago Paternain, Alejandro Ribeiro
Prediction credibility measures, in the form of confidence intervals or probability distributions, are fundamental in statistics and machine learning to characterize model robustne…
Policy Gradient for Continuing Tasks in Non-stationary Markov Decision Processes
Santiago Paternain, Juan Andres Bazerque, Alejandro Ribeiro
Reinforcement learning considers the problem of finding policies that maximize an expected cumulative reward in a Markov decision process with unknown transition probabilities. In…
The empirical duality gap of constrained statistical learning
Luiz F. O. Chamon, Santiago Paternain, Miguel Calvo-Fullana +1
This paper is concerned with the study of constrained statistical learning problems, the unconstrained version of which are at the core of virtually all of modern information proce…
Constrained Reinforcement Learning Has Zero Duality Gap
Santiago Paternain, Luiz F. O. Chamon, Miguel Calvo-Fullana +1
Autonomous agents must often deal with conflicting requirements, such as completing tasks using the least amount of time/energy, learning multiple tasks, or dealing with multiple o…
Sufficiently Accurate Model Learning
Clark Zhang, Arbaaz Khan, Santiago Paternain +1
Modeling how a robot interacts with the environment around it is an important prerequisite for designing control and planning algorithms. In fact, the performance of controllers an…