Summary of Are Expressive Models Truly Necessary For Offline Rl?, by Guan Wang et al.
Are Expressive Models Truly Necessary for Offline RL?
by Guan Wang, Haoyi Niu, Jianxiong Li, Li Jiang, Jianming Hu, Xianyuan Zhan
First submitted to arxiv on: 15 Dec 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 offline reinforcement learning (RL) method called Recursive Skip-Step Planning (RSP), which leverages lightweight models to efficiently capture accurate dynamics across long horizons. Unlike conventional RL methods that rely on large, expressive models, RSP uses shallow 2-layer MLPs and a recursive planning scheme to recursively plan coarse-grained future sub-goals based on current and target information. This approach reduces computation and inference latency while achieving state-of-the-art (SOTA) performances on the D4RL benchmark, particularly in multi-stage long-horizon tasks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Offline reinforcement learning uses goal-conditioned supervised learning to solve sequential modeling problems. A new method called Recursive Skip-Step Planning (RSP) can accurately capture dynamics across long horizons using simple models like shallow 2-layer MLPs. This approach reduces computation and inference latency while achieving better results than large, expressive models on the D4RL benchmark. |
Keywords
» Artificial intelligence » Inference » Reinforcement learning » Supervised
