Haobo Fu
Papers
2
Total Citations
9
H-Index
2
About
Haobo Fu is a rising researcher at the intersection of evolutionary computation and deep reinforcement learning, with a focus on Quality-Diversity (QD) algorithms and multi-task learning. His work on "Multi-objective Optimization-based Selection for Quality-Diversity by Non-surrounded-dominated Sorting" (2023, 5 citations) advances QD algorithms—a subset of evolutionary algorithms that maintain an archive of solutions to generate both high-quality and diverse outputs. By introducing a non-surrounded-dominated sorting mechanism, Fu improves selection pressure within QD archives, enabling more efficient exploration of solution spaces. In parallel, his 2024 paper "Not All Tasks Are Equally Difficult: Multi-Task Deep Reinforcement Learning with Dynamic Depth Routing" (4 citations) tackles the challenge of multi-task RL, where a single policy must handle diverse objectives. Fu proposes a dynamic depth routing approach that adaptively allocates network modules based on task difficulty, enhancing parameter sharing and data efficiency. Though early in his career, Fu’s contributions demonstrate a keen ability to bridge theoretical optimization with practical algorithmic design, offering scalable solutions for complex, real-world problems. His work is particularly impactful for researchers in evolutionary robotics and autonomous systems, where balancing quality and diversity remains a critical challenge.
Research Focus
Key Achievements
Top Papers
- 1
- 2