Summary of System-1.x: Learning to Balance Fast and Slow Planning with Language Models, by Swarnadeep Saha et al.
System-1.x: Learning to Balance Fast and Slow Planning with Language Models
by Swarnadeep Saha, Archiki Prasad, Justin Chih-Yao Chen, Peter Hase, Elias Stengel-Eskin, Mohit Bansal
First submitted to arxiv on: 19 Jul 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: Computation and Language (cs.CL); Machine Learning (cs.LG)
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 System-1.x Planner is a controllable planning framework that uses large language models (LLMs) to solve long-horizon planning problems. It operates in two modes: fast ‘System-1’ mode, which directly generates plans without explicit search or backtracking, and slow ‘System-2’ mode, which searches step-by-step over possible actions. The planner balances between these two modes based on the difficulty of the problem at hand, using a user-specified hybridization factor (x) to govern the mixture. It consists of a controller, System-1 Planner, and System-2 Planner, all fine-tuned on top of a single base LLM requiring only search traces as supervision. The planner outperforms other planners in Maze Navigation and Blocksworld tasks. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The proposed System-1.x Planner is a special kind of computer program that helps us solve problems by making plans. It has two ways to work: one fast way that makes plans quickly, and another slow way that takes its time searching for the best plan. The planner can switch between these two modes depending on how hard the problem is, using a special number (x) to decide. It’s like having a personal assistant that helps you solve problems in different ways! |
