Transformer Dissection: An Unified Understanding for Transformer’s Attention via the Lens of Kernel

Tsai, Yao-Hung Hubert; Bai, Shaojie; Yamada, Makoto; Morency, Louis-Philippe; Salakhutdinov, Ruslan

doi:10.18653/v1/d19-1443

Transformer Dissection: An Unified Understanding for Transformer’s Attention via the Lens of Kernel

Tsai, Yao-Hung Hubert, Bai, Shaojie, Yamada, Makoto · Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP) · 2019