Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper investigates the limitations of large language models (LLMs) in solving planning problems despite their theoretical Turing-completeness. The authors introduce the Chain of Computation (COC) architecture, which integrates a transformer-based language model within an iterative loop and employs a Structured Context Window (SCW) to enhance planning capabilities. The results demonstrate that even small LMs can achieve over 99.89% success rates on planning tasks like BlocksWorld and the Pancake puzzle, while also effectively addressing complex instances of the Tower of Hanoi problem with minimal training data.
Small LLMs can learn effective planning policies and solve complex problems with over 99.89% success rates when using the innovative Chain of Computation architecture.
Large Language Models (LLMs) have had a remarkable impact across many areas of machine learning. However, recent studies have shown that they struggle to reliably solve planning problems. At the same time, theoretical results have shown that transformers, the core architecture underlying modern LLMs, are Turing-complete. In this work, we investigate this apparent gap between the theoretical computational power of LLMs and their empirical planning performance. We propose Chain of Computation (COC), a computational architecture that places a transformer-based LM inside an iterative loop, leveraging its strength as a pattern-matching system. The COC uses a Structured Context Window (SCW) which provides a constant-sized context window with support for choosing which window is used at each planning step. Within this architecture, the LM is able to learn a planning policy, predicts the world model, and performs the arithmetic operations required during planning. We show that, when given an append-only SCW (resembling a Turing Machine tape), even relatively small LMs trained from scratch can learn planning policies and generalize from a small number of training instances within each planning domain, achieving success rates above 99.89\% on BlocksWorld and the Pancake puzzle. Our analysis of failure cases in Tower of Hanoi (TOH) reveals that they arise from arithmetic operations or from encountering previously unseen tokens. We show that COC can solve TOH problem instances with up to 20 disks, requiring over 1 million actions, while requiring substantially less training data by either (1) planning with symbolical support for arithmetic or by (2) using a deterministic pushdown automaton (PDA) formulation for the SCW.