Aditya Mahajan

The standard approach for Partially Observable Markov Decision Processes (POMDPs) is to convert them to a fully observed belief-state MDP. H… (voir plus)owever, the belief state depends on the system model and is therefore not viable in reinforcement learning (RL) settings. A widely used alternative is to use an agent state, which is a model-free, recursively updateable function of the observation history. Examples include frame stacking and recurrent neural networks. Since the agent state is model-free, it is used to adapt standard RL algorithms to POMDPs. However, standard RL algorithms like Q-learning learn a stationary policy. Our main thesis that we illustrate via examples is that because the agent state does not satisfy the Markov property, non-stationary agent-state based policies can outperform stationary ones. To leverage this feature, we propose PASQL (periodic agent-state based Q-learning), which is a variant of agent-state-based Q-learning that learns periodic policies. By combining ideas from periodic Markov chains and stochastic approximation, we rigorously establish that PASQL converges to a cyclic limit and characterize the approximation error of the converged periodic policy. Finally, we present a numerical experiment to highlight the salient features of PASQL and demonstrate the benefit of learning periodic policies over stationary policies.

2024-08-01

EWRL/2024/Workshop (accepté)

openreview.net

On learning history-based policies for controlling Markov decision processes

Gandharv Patil

Doina Precup

2024-04-18

Proceedings of The 27th International Conference on Artificial Intelligence and Statistics (publié)

openreview.net

Model approximation in MDPs with unbounded per-step cost

Berk Bozkurt

Ashutosh Nayyar

Yi Ouyang

We consider the problem of designing a control policy for an infinite-horizon discounted cost Markov decision process …

2024-02-13

ArXiv (prépublication)

Bridging State and History Representations: Understanding Self-Predictive RL

Tianwei Ni

Benjamin Eysenbach

Erfan SeyedSalehi

Michel Ma

Clement Gehring

Pierre-Luc Bacon

Representations are at the core of all deep reinforcement learning (RL) methods for both Markov decision processes (MDPs) and partially obse… (voir plus)rvable Markov decision processes (POMDPs). Many representation learning methods and theoretical frameworks have been developed to understand what constitutes an effective representation. However, the relationships between these methods and the shared properties among them remain unclear. In this paper, we show that many of these seemingly distinct methods and frameworks for state and history abstractions are, in fact, based on a common idea of self-predictive abstraction. Furthermore, we provide theoretical insights into the widely adopted objectives and optimization, such as the stop-gradient technique, in learning self-predictive representations. These findings together yield a minimalist algorithm to learn self-predictive representations for states and histories. We validate our theories by applying our algorithm to standard MDPs, MDPs with distractors, and POMDPs with sparse rewards. These findings culminate in a set of preliminary guidelines for RL practitioners.

2024-01-16

ICLR.cc/2024/Conference (poster)

openreview.net

On learning Whittle index policy for restless bandits with scalable regret

Nima Akbarzadeh

Reinforcement learning is an attractive approach to learn good resource allocation and scheduling policies based on data when the system mod… (voir plus)el is unknown. However, the cumulative regret of most RL algorithms scales as ˜ O(S

2024-01-01

IEEE Transactions on Control of Network Systems (publié)

Strong Consistency and Rate of Convergence of Switched Least Squares System Identification for Autonomous Markov Jump Linear Systems

Borna Sayedana

Mohammad Afshari

Peter E. Caines

In this paper, we investigate the problem of system identification for autonomous Markov jump linear systems (MJS) with complete state obser… (voir plus)vations. We propose switched least squares method for identification of MJS, show that this method is strongly consistent, and derive data-dependent and data-independent rates of convergence. In particular, our data-independent rate of convergence shows that, almost surely, the system identification error is

2024-01-01

IEEE Transactions on Automatic Control (publié)

Two Families of Indexable Partially Observable Restless Bandits and Whittle Index Computation

Nima Akbarzadeh

2024-01-01

Performance Evaluation (publié)

Asymmetric Actor-Critic with Approximate Information State

Amit Sinha

Reinforcement learning (RL) for partially observable Markov decision processes (POMDPs) is a challenging problem because decisions need to b… (voir plus)e made based on the entire history of observations and actions. However, in several scenarios, state information is available during the training phase. We are interested in exploiting the availability of this state information during the training phase to efficiently learn a history-based policy using RL. Specifically, we consider actor-critic algorithms, where the actor uses only the history information but the critic uses both history and state. Such algorithms are called asymmetric actor-critic, to highlight the fact that the actor and critic have asymmetric information. Motivated by the recent success of using representation losses in RL for POMDPs [1], we derive similar theoretical results for the asymmetric actor-critic case and evaluate the effectiveness of adding such auxiliary losses in experiments. In particular, we learn a history representation-called an approximate information state (AIS)-and bound the performance loss when acting using AIS.

2023-12-13

IEEE Conference on Decision and Control (publié)

Relative Almost Sure Regret Bounds for Certainty Equivalence Control of Markov Jump Systems

Borna Sayedana

Mohammad Afshari

Peter E. Caines

In this paper, we consider learning and control problem in an unknown Markov jump linear system (MJLS) with perfect state observations. We f… (voir plus)irst establish a generic upper bound on regret for any learning based algorithm. We then propose a certainty equivalence-based learning alagrithm and show that this algorithm achieves a regret of

2023-12-13

2023 62nd IEEE Conference on Decision and Control (CDC) (publié)

Weighted-Norm Bounds on Model Approximation in MDPs with Unbounded Per-Step Cost

Berk Bozkurt

Ashutosh Nayyar

Yi Ouyang

We consider the problem of designing a control policy for an infinite-horizon discounted cost Markov Decision Process (MDP) …

2023-12-13

IEEE Conference on Decision and Control (publié)