Search Results for author: Xiangkun He

Found 4 papers, 0 papers with code

Goal-guided Transformer-enabled Reinforcement Learning for Efficient Autonomous Navigation

no code implementations1 Jan 2023 Wenhui Huang, Yanxin Zhou, Xiangkun He, Chen Lv

Despite some successful applications of goal-driven navigation, existing deep reinforcement learning-based approaches notoriously suffers from poor data efficiency issue.

Autonomous Navigation Decision Making +2

Safe Decision-making for Lane-change of Autonomous Vehicles via Human Demonstration-aided Reinforcement Learning

no code implementations1 Jul 2022 Jingda Wu, Wenhui Huang, Niels de Boer, Yanghui Mo, Xiangkun He, Chen Lv

Decisions made by human subjects in a driving simulator are treated as safe demonstrations, which are stored into the replay buffer and then utilized to enhance the training process of RL.

Autonomous Driving Decision Making +1

Sampling Efficient Deep Reinforcement Learning through Preference-Guided Stochastic Exploration

no code implementations20 Jun 2022 Wenhui Huang, Cong Zhang, Jingda Wu, Xiangkun He, Jie Zhang, Chen Lv

We theoretically prove that the policy improvement theorem holds for the preference-guided $\epsilon$-greedy policy and experimentally show that the inferred action preference distribution aligns with the landscape of corresponding Q-values.

Q-Learning reinforcement-learning +1

Approximating Pareto Frontier through Bayesian-optimization-directed Robust Multi-objective Reinforcement Learning

no code implementations1 Jan 2021 Xiangkun He, Jianye Hao, Dong Li, Bin Wang, Wulong Liu

Thirdly, the agent’s learning process is regarded as a black-box, and the comprehensive metric we proposed is computed after each episode of training, then a Bayesian optimization (BO) algorithm is adopted to guide the agent to evolve towards improving the quality of the approximated Pareto frontier.

Multi-Objective Reinforcement Learning reinforcement-learning

Cannot find the paper you are looking for? You can Submit a new open access paper.