Title: Linformer: Self-Attention with Linear Complexity
Authors: Sinong Wang, Belinda Z. Li, Madian Khabsa, Han Fang, Hao Ma
Published: 8th June 2020 (Monday) @ 17:37:52
Link: http://arxiv.org/abs/2006.04768v3
Abstract
Large transformer models have shown extraordinary success in achieving state-of-the-art results in many natural language processing applications. However, training and deploying these models can be prohibitively costly for long sequences, as the standard self-attention mechanism of the Transformer uses time and space with respect to sequence length. In this paper, we demonstrate that the self-attention mechanism can be approximated by a low-rank matrix. We further exploit this finding to propose a new self-attention mechanism, which reduces the overall self-attention complexity from to in both time and space. The resulting linear transformer, the \textit{Linformer}, performs on par with standard Transformer models, while being much more memory- and time-efficient.