Latent Landmark Graph for Efficient Exploration-exploitation Balance in Hierarchical Reinforcement Learning

Q Zhang, H Zhang, D **ng, B Xu - Machine Intelligence Research, 2025 - Springer
Goal-conditioned hierarchical reinforcement learning (GCHRL) decomposes the desired
goal into subgoals and conducts exploration and exploitation in the subgoal space. Its …

Decoding BatchNorm statistics via anchors pool for data-free models based on continual learning

X Li, W Wang, G Xu - Neural Computing and Applications, 2024 - Springer
Generating high-quality samples reversely from existing models is a significant technique in
continual learning and knowledge distillation. Existing approaches either fail to generate …

Trajectory Progress-Based Prioritizing and Intrinsic Reward Mechanism for Robust Training of Robotic Manipulations

W Liang, Y Liu, J Wang, ZX Yang - IEEE Transactions on …, 2024 - ieeexplore.ieee.org
Training robots by model-free deep reinforcement learning (DRL) to carry out robotic
manipulation tasks without sufficient successful experiences is challenging. Hindsight …

Natural Mitigation of Catastrophic Interference: Continual Learning in Power-Law Learning Environments

A Gandhi, RS Shah, V Marupudi… - arxiv preprint arxiv …, 2024 - ebooks.iospress.nl
Neural networks often suffer from catastrophic interference (CI): performance on previously
learned tasks drops off significantly when learning a new task. This contrasts strongly with …