Qi Zhu

TIMESLIVER : SYMBOLIC-LINEAR DECOMPOSITION FOR EXPLAINABLE TIME SERIES CLASSIFICATION

ODESteer: A Unified ODE-Based Steering Framework for LLM Alignment

Efficient Multi-Step Reinforcement Learning with Expectation-Maximization Bootstrapping

Belief-Based Offline Reinforcement Learning for Delay-Robust Policy Optimization

SENTINEL: A Multi-Level Formal Framework for Safety Evaluation of LLM-based Embodied Agents

MAESTRO : Adaptive Sparse Attention and Robust Learning for Multimodal Dynamic Time Series

On Large Language Model Continual Unlearning

Directly Forecasting Belief for Reinforcement Learning with Delays

Can We Trust Embodied Agents? Exploring Backdoor Attacks against Embodied LLM-Based Decision-Making Systems

Model-Enhanced Adversarial Inverse Reinforcement Learning with Model Estimation Reward Shaping in Stochastic Environments

Phase-Driven Domain Generalizable Learning For Nonstationary Time Series Classification