Publications

This paper presents a novel approach to directly forecast beliefs in reinforcement learning with observation delays, improving upon traditional methods by incorporating predictive capabilities into the learning process.

Variational Delayed Policy Optimization (VDPO) reformulates delayed RL as a variational inference problem, which is further modelled as a two-step iterative optimization problem, where the first step is TD learning in the delay-free environment with a small state space, and the second step is behaviour cloning which can be addressed much more efficiently than TD learning.

This paper presents a novel approach to trajectory generation and prediction that incorporates kinematic constraints through latent stochastic differential equations, enabling more realistic and physically-consistent motion planning.

Auxiliary-Delayed Reinforcement Learning (AD-RL) leverages an auxiliary short-delayed task to accelerate the learning on a strongly delayed task without compromising the performance in stochastic environments.

In this paper, we propose a novel pixel-observation safe RL algorithm that efficiently encodes state-wise safety constraints with unknown hazard regions through the introduction of a latent barrier function learning mechanism.

A safe RL approach that can jointly learn the environment and optimize the control policy, while effectively avoiding unsafe regions with safety probability optimization.

A framework that jointly conducts reinforcement learning and formal verification by formulating and solving a novel bilevel optimization problem, which is end-to-end differentiable by the gradients from the value function and certificates formulated by linear programs and semi-definite programs.

MicroFluID is a novel RFID artifact based on a multiple-chip structure and microfluidic switches, which informs the input state by directly reading variable ID information instead of retrieving primitive signals.

RElectrode is a reconfigurable electrode using a microfluidic technique that can change the geometry and material properties of the electrode to satisfy the needs for sensing a variety of different types of user input through touch/touchless gestures, pressure, temperature, and distinguish between different types of objects or liquids.

This paper introduces Shop-R1, a novel reinforcement learning framework aimed at enhancing the reasoning ability of LLMs for simulation of real human behavior in online shopping environments through a two-stage approach with distinct reward signals.

This paper explores the integration of Large Language Models (LLMs) into autonomous driving systems, leveraging their robust common-sense knowledge and reasoning abilities to enhance driving performance and safety in long-tail unforeseen scenarios.