田霞

个人信息

参与实验室科研项目

人机混合智能系统双层智能测试评估技术研究

复杂环境下非完全信息博弈决策的智能基础模型研究

研究课题

基于指导水平的人在环深度强化学习方法研究

学术成果

共撰写/参与撰写专利 1 项，录用/发表论文 2 篇，投出待录用论文1篇。 联培学生可能有其他不在此展示的论文/专利。

patent

1. 眼科病人信息录入软件V1.0 康宇, 田霞, 董凯, 鲁理, 夏睿钰, 赵云波, 刘斌琨, and 李晓蒙 2023 [pdf]

Journal Articles

1. A Dual Confidence Evaluation-Based Shared Control Approach for Human-Machine Collaboration Yaqing Zhao, and Yun-Bo Zhao 2026 [Abs] [pdf]

   In recent years, shared control has received widespread attention for its potential to enhance humanmachine collaboration (HMC) performance. The core challenge lies in how to reasonably allocate control authority between humans and machines. However, existing rule-based or single-confidencebased allocation approaches often suffer from limited flexibility, high sensitivity to environmental changes, and complex mappings from decision signals, making them less effective in meeting the demands of dynamic tasks. To address these limitations, this paper proposes a Dual Confidence-based Shared Control (DC-SC) approach. First, human and machine decision confidence evaluation models are established to quantitatively describe the performance of both parties in the decision-making process. Then, a learnable shared control policy is designed based on these confidence evaluations, and the control authority is dynamically optimized in real time through environmental feedback within the policy gradient optimization framework, aiming to achieve optimal system performance in humanmachine collaboration. In addition, the key properties of the proposed confidence evaluation models are systematically analyzed from a theoretical perspective, further validating the rationality and reliability of the model design. Finally, experiments are conducted in both discrete and continuous action tasks. The results show that the proposed DC-SC approach outperforms both classical reinforcement learning-based machine control approach (A2C/DDPG) and a representative uncertainty-aware shared control approach (HULA) in terms of policy performance and system safety.

Book Chapters

1. Uncertainty-Based Dynamic Weighted Experience Replay for Human-in-the-Loop Deep Reinforcement Learning Xia Tian, Yu Kang, Yunbo Zhao, Yaqing Zhou, and Pengfei Li In Frontiers in Artificial Intelligence and Applications 2025 [Abs] [doi] [pdf]

   Human-in-the-loop reinforcement learning (HIRL) enhances sampling efficiency in deep reinforcement learning by incorporating human expertise and experience into the training process. However, HIRL methods still heavily depend on expert guidance, which is a key factor limiting their further development and largescale application. In this paper, an uncertainty-based dynamic weighted experience replay approach (UDWER) is proposed to solve the above problem. Our approach enables the algorithm to detect decision uncertainty, triggering human intervention only when uncertainty exceeds a threshold. This reduces the need for continuous human supervision. Additionally, we design a dynamic experience replay mechanism that prioritizes machine self-exploration and human-guided samples with different weights based on decision uncertainty. We also provide a theoretical derivation and related discussion. Experiments in the Lunar Lander environment demonstrate improved sampling efficiency and reduced reliance on human guidance.

博客文章

学位论文

毕业去向

宝武钢铁股份有限公司， 智能制造工程师