Short-Long Convolutions Help Hardware-Efficient Linear Attention to Focus on Long Sequences

TL;DR

This paper introduces CHELA, a hybrid approach combining short-long convolutions with hardware-efficient linear attention, enabling effective long sequence processing with real linear complexity and improved stability.

Contribution

It proposes a novel hybrid model that replaces SSMs with convolutions and implements linear attention efficiently for long sequences, addressing hardware and stability issues.

Findings

Outperforms existing methods on Long Range Arena benchmark

Achieves real linear complexity in long sequence processing

Demonstrates effectiveness on language modeling tasks

Abstract

To mitigate the computational complexity in the self-attention mechanism on long sequences, linear attention utilizes computation tricks to achieve linear complexity, while state space models (SSMs) popularize a favorable practice of using non-data-dependent memory pattern, i.e., emphasize the near and neglect the distant, to processing sequences. Recent studies have shown the priorities by combining them as one. However, the efficiency of linear attention remains only at the theoretical level in a causal setting, and SSMs require various designed constraints to operate effectively on specific data. Therefore, in order to unveil the true power of the hybrid design, the following two issues need to be addressed: (1) hardware-efficient implementation for linear attention and (2) stabilization of SSMs. To achieve this, we leverage the thought of tiling and hierarchy to propose CHELA (short-long Convolutions with Hardware-Efficient Linear Attention), which replaces SSMs with short-long convolutions and implements linear attention in a divide-and-conquer manner. This approach enjoys global abstraction and data-dependent selection from stable SSM and linear attention while maintaining real linear complexity. Our comprehensive experiments on the Long Range Arena benchmark and language modeling tasks demonstrate the effectiveness of the proposed method.