Summary of Probabilistic Subgoal Representations For Hierarchical Reinforcement Learning, by Vivienne Huiling Wang et al.
Probabilistic Subgoal Representations for Hierarchical Reinforcement learning
by Vivienne Huiling Wang, Tinghuai Wang, Wenyan Yang, Joni-Kristian Kämäräinen, Joni Pajarinen
First submitted to arxiv on: 24 Jun 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper proposes a novel approach to goal-conditioned hierarchical reinforcement learning (HRL) by introducing probabilistic subgoal representations using Gaussian Processes (GPs). This allows for adaptive memory and coping with stochastic uncertainties. The method learns a posterior distribution over the subgoal representation functions, exploiting long-range correlation in the state space through learnable kernels. The authors also propose a novel learning objective to facilitate simultaneous learning of probabilistic subgoal representations and policies within a unified framework. Experiments show that the approach outperforms state-of-the-art baselines in standard benchmarks and diverse reward conditions. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary In goal-conditioned hierarchical reinforcement learning, a high-level policy specifies a subgoal for a low-level policy to reach. The new method uses Gaussian Processes (GPs) to learn a probabilistic subgoal representation that helps the agent cope with uncertainty. This allows it to remember long-term plans and make better decisions. The authors also developed a new way to train both the subgoal representations and policies together, which improves performance. The approach does well in standard tests and ones with unpredictable outcomes. |
Keywords
* Artificial intelligence * Reinforcement learning
