Linear Video Transformer with Feature Fixation

Kaiyue Lu, ZeXiang Liu, Jianyuan Wang, Weixuan Sun, Zhen Qin, Dong Li, Xuyang Shen, Hui Deng, Xiaodong Han, Yuchao Dai, Yiran Zhong

Abstract: Vision Transformers have achieved impressive performance in video classification, while suffering from the quadratic complexity caused by the Softmax attention mechanism. Some studies alleviate the computational costs by reducing the number of tokens attended in attention calculation, but the complexity is still quadratic. Another promising way is to replace Softmax attention with linear attention, which owns linear complexity but presents a clear performance drop. We find that such a drop in linear attention results from the lack of attention concentration to critical features. Therefore, we propose a feature fixation module to reweight feature importance of the query and key prior to computing linear attention. Specifically, we regard the query, key, and value as latent representations of the input token, and learn the feature fixation ratio by aggregating Query-Key-Value information. This is beneficial for measuring the feature importance comprehensively. Furthermore, we improve the feature fixation by neighborhood association, which leverages additional guidance from spatial and temporal neighboring tokens. Our proposed method significantly improves the linear attention baseline, and achieves state-of-the-art performance among linear video Transformers on three popular video classification benchmarks. Our performance is even comparable to some quadratic Transformers with fewer parameters and higher efficiency.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Applications (eg, speech processing, computer vision, NLP)