Efficient Attention: Attention with Linear Complexities
This addresses the resource inefficiency problem for researchers and practitioners in computer vision and NLP, enabling broader integration of attention in complex models, though it is incremental as it builds on existing attention paradigms.
The paper tackled the quadratic memory and computational costs of dot-product attention, which limits its use on high-resolution inputs, by proposing an efficient attention mechanism with linear complexities, leading to significant performance boosts on MS-COCO 2017 and state-of-the-art accuracies on the Scene Flow dataset.
Dot-product attention has wide applications in computer vision and natural language processing. However, its memory and computational costs grow quadratically with the input size. Such growth prohibits its application on high-resolution inputs. To remedy this drawback, this paper proposes a novel efficient attention mechanism equivalent to dot-product attention but with substantially less memory and computational costs. Its resource efficiency allows more widespread and flexible integration of attention modules into a network, which leads to better accuracies. Empirical evaluations demonstrated the effectiveness of its advantages. Efficient attention modules brought significant performance boosts to object detectors and instance segmenters on MS-COCO 2017. Further, the resource efficiency democratizes attention to complex models, where high costs prohibit the use of dot-product attention. As an exemplar, a model with efficient attention achieved state-of-the-art accuracies for stereo depth estimation on the Scene Flow dataset. Code is available at https://github.com/cmsflash/efficient-attention.