Chain of Thought Empowers Transformers to Solve Inherently Serial Problems

• URL: http://arxiv.org/abs/2402.12875v4
• Date: Sat, 21 Sep 2024 06:48:45 GMT
• Title: Chain of Thought Empowers Transformers to Solve Inherently Serial Problems
• Authors: Zhiyuan Li, Hong Liu, Denny Zhou, Tengyu Ma,
• Abstract summary: Chain of thought (CoT) is a highly effective method to improve the accuracy of large language models (LLMs) on arithmetics and symbolic reasoning tasks. This work provides a theoretical understanding of the power of CoT for decoder-only transformers through the lens of expressiveness.
• Score: 57.58801785642868
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Instructing the model to generate a sequence of intermediate steps, a.k.a., a chain of thought (CoT), is a highly effective method to improve the accuracy of large language models (LLMs) on arithmetics and symbolic reasoning tasks. However, the mechanism behind CoT remains unclear. This work provides a theoretical understanding of the power of CoT for decoder-only transformers through the lens of expressiveness. Conceptually, CoT empowers the model with the ability to perform inherently serial computation, which is otherwise lacking in transformers, especially when depth is low. Given input length $n$, previous works have shown that constant-depth transformers with finite precision $\mathsf{poly}(n)$ embedding size can only solve problems in $\mathsf{TC}^0$ without CoT. We first show an even tighter expressiveness upper bound for constant-depth transformers with constant-bit precision, which can only solve problems in $\mathsf{AC}^0$, a proper subset of $ \mathsf{TC}^0$. However, with $T$ steps of CoT, constant-depth transformers using constant-bit precision and $O(\log n)$ embedding size can solve any problem solvable by boolean circuits of size $T$. Empirically, enabling CoT dramatically improves the accuracy for tasks that are hard for parallel computation, including the composition of permutation groups, iterated squaring, and circuit value problems, especially for low-depth transformers.

Related papers

• Lower Bounds for Chain-of-Thought Reasoning in Hard-Attention Transformers [5.4649464326326]
  Chain-of-thought reasoning and scratchpads have emerged as critical tools for enhancing the computational capabilities of transformers. In this work, we initiate the study of systematic lower bounds for the number of CoT steps across different algorithmic problems.
  arXiv  Detail & Related papers  (2025-02-04T15:14:01Z)
• Theoretical limitations of multi-layer Transformer [14.63344366356708]
  We prove the first $textitunconditional$ lower bound against multi-layer decoder-only transformers. We also introduce a new proof technique that finds a certain $textitindistinguishable$ $textitde$ all possible inputs. We believe our new communication model and proof technique will be helpful to further understand the computational power of transformers.
  arXiv  Detail & Related papers  (2024-12-04T02:37:31Z)
• Circuit Complexity Bounds for RoPE-based Transformer Architecture [25.2590541420499]
  Empirical evidence suggests that $mathsfRoPE$-based Transformer architectures demonstrate greater generalization capabilities than conventional Transformer models. We show that unless $mathsfTC0 = mathsfNC1$, a $mathsfRoPE$-based Transformer with $mathrmpoly(n)$-precision, $O(1)$ layers, hidden dimension $d leq O(n)$ cannot solve the Arithmetic formula evaluation problem.
  arXiv  Detail & Related papers  (2024-11-12T07:24:41Z)
• On the Role of Depth and Looping for In-Context Learning with Task Diversity [69.4145579827826]
  We study in-context learning for linear regression with diverse tasks. We show that multilayer Transformers are not robust to even distributional shifts as small as $O(e-L)$ in Wasserstein distance.
  arXiv  Detail & Related papers  (2024-10-29T03:27:56Z)
• Transformers Learn Shortcuts to Automata [52.015990420075944]
  We find that a low-depth Transformer can represent the computations of any finite-state automaton. We show that a Transformer with $O(log T)$ layers can exactly replicate the computation of an automaton on an input sequence of length $T$. We further investigate the brittleness of these solutions and propose potential mitigations.
  arXiv  Detail & Related papers  (2022-10-19T17:45:48Z)
• The Parallelism Tradeoff: Limitations of Log-Precision Transformers [29.716269397142973]
  We prove that transformers whose arithmetic precision is logarithmic in the number of input tokens can be simulated by constant-depth logspace-uniform threshold circuits. This provides insight on the power of transformers using known results in complexity theory.
  arXiv  Detail & Related papers  (2022-07-02T03:49:34Z)
• What Dense Graph Do You Need for Self-Attention? [73.82686008622596]
  We present Hypercube Transformer, a sparse Transformer that models token interactions in a hypercube and shows comparable or even better results with vanilla Transformer. Experiments on tasks requiring various sequence lengths lay validation for our graph function well.
  arXiv  Detail & Related papers  (2022-05-27T14:36:55Z)
• Stable, Fast and Accurate: Kernelized Attention with Relative Positional Encoding [63.539333383965726]
  We propose a novel way to accelerate attention calculation for Transformers with relative positional encoding (RPE) Based upon the observation that relative positional encoding forms a Toeplitz matrix, we mathematically show that kernelized attention with RPE can be calculated efficiently using Fast Fourier Transform (FFT)
  arXiv  Detail & Related papers  (2021-06-23T17:51:26Z)
• $O(n)$ Connections are Expressive Enough: Universal Approximability of Sparse Transformers [71.31712741938837]
  We show that sparse Transformers with only $O(n)$ connections per attention layer can approximate the same function class as the dense model with $n2$ connections. We also present experiments comparing different patterns/levels of sparsity on standard NLP tasks.
  arXiv  Detail & Related papers  (2020-06-08T18:30:12Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.