Summary of Language Models Can Infer Action Semantics For Symbolic Planners From Environment Feedback, by Wang Zhu et al.
Language Models can Infer Action Semantics for Symbolic Planners from Environment Feedback
by Wang Zhu, Ishika Singh, Robin Jia, Jesse Thomason
First submitted to arxiv on: 4 Jun 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: Computation and Language (cs.CL); Robotics (cs.RO)
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 proposed Predicting Semantics of Actions with Language Models (PSALM) approach combines symbolic planners and Large Language Models (LLMs) to automatically learn domain-specific action semantics. PSALM leverages the strengths of both methods by repeatedly proposing and executing plans, using the LLM to partially generate plans and infer action semantics based on execution outcomes. This iterative process maintains a belief over possible action semantics until a goal state is reached. The approach is evaluated on 7 environments, demonstrating significant improvements in plan success rate and environment exploration efficiency compared to prior work. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary PSALM uses two types of models: symbolic planners and large language models (LLMs). Symbolic planners are good at understanding rules and making decisions based on them. LLMs can generate text but struggle with decision-making. PSALM combines these strengths by letting the LLM help make decisions and figure out what actions to take, while also using the planner to ensure those actions are correct. This way, PSALM can create better plans that work in real-world situations. |
Keywords
» Artificial intelligence » Semantics
