Summary of Autosat: Automatically Optimize Sat Solvers Via Large Language Models, by Yiwen Sun et al.
AutoSAT: Automatically Optimize SAT Solvers via Large Language Models
by Yiwen Sun, Furong Ye, Xianyin Zhang, Shiyu Huang, Bingzhen Zhang, Ke Wei, Shaowei Cai
First submitted to arxiv on: 16 Feb 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: None
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 This paper introduces a framework called AutoSAT that automatically optimizes heuristics in conflict-driven clause learning (CDCL) solvers for solving satisfiability problems. CDCL solvers, such as MiniSat and Kissat, rely on various heuristics to improve their performance, but selecting the right heuristic can be time-consuming and requires expertise. The authors propose using large language models (LLMs) to generate new efficient heuristics, which can enhance the performance of CDCL solvers. They employ strategies like greedy hill climbing and evolutionary algorithms to guide LLMs in searching for better heuristics. Experimental results show that LLMs can generally improve the performance of CDCL solvers, with a realized version of AutoSAT outperforming MiniSat on 9 out of 12 datasets and even surpassing the state-of-the-art hybrid solver Kissat on 4 datasets. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper makes it easier to make computer programs solve tricky math problems called satisfiability problems. Right now, people have to use special tricks (called heuristics) to help these programs work better, but finding the right trick can be hard and takes a lot of time and expertise. The authors came up with an idea to use super-powerful language models to find new and better tricks for solving math problems. They tested their idea and found that it worked well, making the computer programs solve problems faster and better than before. |
