How Sparse Attention Approximates Exact Attention? Your Attention is Naturally $n^C$-Sparse

TL;DR

This paper provides a theoretical analysis of sparse attention, revealing it is inherently $n^C$-sparse and establishing conditions under which sparse attention approximates exact attention effectively, guiding more efficient model design.

Contribution

It introduces a theoretical framework that explains the inherent sparsity of standard attention and proposes adaptive strategies for improved sparse attention methods.

Findings

Attention is $n^{C}$-sparse, with only the largest $ ext{Omega}(n^{C})$ entries needed.

Stable $o( ext{log}(n))$-sparse attention cannot fully approximate attention due to persistent error.

Adaptive window size strategies outperform fixed ones in accuracy and efficiency for flexible context lengths.

Abstract

Sparse Attention is a technique that approximates standard attention computation with sub-quadratic complexity. This is achieved by selectively ignoring smaller entries in the attention matrix during the softmax function computation. Variations of this technique, such as pruning KV cache, sparsity-based fast attention, and Sparse Transformer, have been extensively utilized for efficient Large Language Models (LLMs) deployment. Despite its widespread use, a theoretical understanding of the conditions under which sparse attention performs on par with traditional attention remains elusive. This work aims to $\textbf{bridge this gap by examining the inherent sparsity of standard attention processes}$. Our theoretical framework reveals several brand-new key insights: $\bullet$ Attention is $n^{C}$-sparse, implying that considering only the largest $\Omega(n^{C})$ entries out of all $n$ entries is sufficient for sparse attention to approximate the exact attention matrix with decreasing loss. Here, $n$ represents the input length and $C \in (0, 1)$ is a constant. $\bullet$ Stable $o(\log(n))$-sparse attention, which approximates attention computation with $\log(n)$ or fewer entries, may not be feasible since the error will persist at a minimum of $O(1)$. $\bullet$ An adaptive strategy ($\alpha \cdot n^C, \alpha \in \mathbb{R}$) for the window size of efficient attention methods rather than a fixed one is guaranteed to perform more accurately and efficiently in a task for inference on flexible context lengths.