Summary of Learning Off-policy with Model-based Intrinsic Motivation For Active Online Exploration, by Yibo Wang et al.
Learning Off-policy with Model-based Intrinsic Motivation For Active Online Exploration
by Yibo Wang, Jiang Zhao
First submitted to arxiv on: 31 Mar 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 an RL algorithm that achieves sample-efficient exploration in continuous control tasks by incorporating a predictive model and off-policy learning elements. The algorithm uses an online planner enhanced by a novelty-aware terminal value function for sample collection. It also derives an intrinsic reward from the forward predictive error within a latent state space, which establishes a connection to model uncertainty and helps the agent overcome performance gaps. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper aims to improve exploration ability in sample-efficient RL, building on previous advancements in deep reinforcement learning. The algorithm is designed to efficiently collect samples while exploring the environment, using a novelty-aware terminal value function to guide the process. The results show competitive or even superior performance compared to prior works, especially in sparse reward cases. |
Keywords
* Artificial intelligence * Reinforcement learning
