Xiaoming Duan

Yash Paliwal, Rajarshi Roy, Jean-Raphaël Gaglione et al.

We study a class of reinforcement learning (RL) tasks where the objective of the agent is to accomplish temporally extended goals. In this setting, a common approach is to represent the tasks as deterministic finite automata (DFA) and integrate them into the state-space for RL algorithms. However, while these machines model the reward function, they often overlook the causal knowledge about the environment. To address this limitation, we propose the Temporal-Logic-based Causal Diagram (TL-CD) in RL, which captures the temporal causal relationships between different properties of the environment. We exploit the TL-CD to devise an RL algorithm in which an agent requires significantly less exploration of the environment. To this end, based on a TL-CD and a task DFA, we identify configurations where the agent can determine the expected rewards early during an exploration. Through a series of case studies, we demonstrate the benefits of using TL-CDs, particularly the faster convergence of the algorithm to an optimal policy due to reduced exploration of the environment.

Haoxuan Pan, Deheng Ye, Xiaoming Duan et al.

We revisit the estimation bias in policy gradients for the discounted episodic Markov decision process (MDP) from Deep Reinforcement Learning (DRL) perspective. The objective is formulated theoretically as the expected returns discounted over the time horizon. One of the major policy gradient biases is the state distribution shift: the state distribution used to estimate the gradients differs from the theoretical formulation in that it does not take into account the discount factor. Existing discussion of the influence of this bias was limited to the tabular and softmax cases in the literature. Therefore, in this paper, we extend it to the DRL setting where the policy is parameterized and demonstrate how this bias can lead to suboptimal policies theoretically. We then discuss why the empirically inaccurate implementations with shifted state distribution can still be effective. We show that, despite such state distribution shift, the policy gradient estimation bias can be reduced in the following three ways: 1) a small learning rate; 2) an adaptive-learning-rate-based optimizer; and 3) KL regularization. Specifically, we show that a smaller learning rate, or, an adaptive learning rate, such as that used by Adam and RSMProp optimizers, makes the policy optimization robust to the bias. We further draw connections between optimizers and the optimization regularization to show that both the KL and the reverse KL regularization can significantly rectify this bias. Moreover, we provide extensive experiments on continuous control tasks to support our analysis. Our paper sheds light on how successful PG algorithms optimize policies in the DRL setting, and contributes insights into the practical issues in DRL.

Chendi Qu, Jianping He, Xiaoming Duan et al.

Inverse reinforcement learning (IRL) usually assumes the reward function model is pre-specified as a weighted sum of features and estimates the weighting parameters only. However, how to select features and determine a proper reward model is nontrivial and experience-dependent. A simplistic model is less likely to contain the ideal reward function, while a model with high complexity leads to substantial computation cost and potential overfitting. This paper addresses this trade-off in the model selection for IRL problems by introducing the structural risk minimization (SRM) framework from statistical learning. SRM selects an optimal reward function class from a hypothesis set minimizing both estimation error and model complexity. To formulate an SRM scheme for IRL, we estimate the policy gradient from given demonstration as the empirical risk, and establish the upper bound of Rademacher complexity as the model penalty of hypothesis function classes. The SRM learning guarantee is further presented. In particular, we provide the explicit form for the linear weighted sum setting. Simulations demonstrate the performance and efficiency of our algorithm.

Zhe Xu, Xiaoming Duan

Pandemics can bring a range of devastating consequences to public health and the world economy. Identifying the most effective control strategies has been the imperative task all around the world. Various public health control strategies have been proposed and tested against pandemic diseases (e.g., COVID-19). We study two specific pandemic control models: the susceptible, exposed, infectious, recovered (SEIR) model with vaccination control; and the SEIR model with shield immunity control. We express the pandemic control requirement in metric temporal logic (MTL) formulas. We then develop an iterative approach for synthesizing the optimal control strategies with MTL specifications. We provide simulation results in two different scenarios for robust control of the COVID-19 pandemic: one for vaccination control, and another for shield immunity control, with the model parameters estimated from data in Lombardy, Italy. The results show that the proposed synthesis approach can generate control inputs such that the time-varying numbers of individuals in each category (e.g., infectious, immune) satisfy the MTL specifications with robustness against initial state and parameter uncertainties.