no code implementations • 4 May 2024 • Jaeyoon Kim, Junyu Xuan, Christy Liang, Farookh Hussain
Our suggested model is termed Non-Monolithic unsupervised Pre-training with Successor features (NMPS), which improves the performance of the original monolithic exploration method of pre-training with SFs.
1 code implementation • 2 May 2023 • Jaeyoon Kim, Junyu Xuan, Christy Liang, Farookh Hussain
The issue of `when' of a monolithic exploration in the usual RL exploration behaviour binds an exploratory action to an exploitational action of an agent.
no code implementations • 20 Nov 2021 • Zhixiong Yue, Feiyang Ye, Yu Zhang, Christy Liang, Ivor W. Tsang
We theoretically study the safeness of both learning strategies in the DSMTL model to show that the proposed methods can achieve some versions of safe multi-task learning.
1 code implementation • 17 Jul 2021 • Jaeyoon Kim, Junyu Xuan, Christy Liang, Farookh Hussain
We propose a novel HRL model supporting the optimal level synchronization using the off-policy correction technique with a deep generative model.
Hierarchical Reinforcement Learning reinforcement-learning +1