Summary of Can Transformers Reason Logically? a Study in Sat Solving, by Leyan Pan et al.
Can Transformers Reason Logically? A Study in SAT Solving
by Leyan Pan, Vijay Ganesh, Jacob Abernethy, Chris Esposo, Wenke Lee
First submitted to arxiv on: 9 Oct 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Logic in Computer Science (cs.LO)
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 investigates the logical reasoning capabilities of decoder-only Transformers in the context of the boolean satisfiability (SAT) problem. It proves that these models can decide 3-SAT using backtracking and deduction via Chain-of-Thought (CoT), showcasing their ability to reason logically. The authors then implement this construction as a PyTorch model, demonstrating its correctness through empirical experiments. Furthermore, they explore whether the model can be trained directly from algorithmic traces (“reasoning paths”) from their theoretical construction, achieving strong out-of-distribution generalization on problem sizes seen during training. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper looks at how well computers can reason using a special kind of AI called decoder-only Transformers. It shows that these models can solve certain types of logic problems really well. The authors then create a real computer program based on this idea and test it to make sure it works correctly. They also try to teach the model to solve logic problems by showing it how other computers solve them, which helps it learn quickly. |
Keywords
» Artificial intelligence » Decoder » Generalization
