#ThisWeekInTheLab Jonathan Jakob talked about follow-up work on our paper on SAM-kNN regression for Online Learning -- see https://t.co/qJcni6FzLI In particular, he investigated how to use metric learning for improved interpretability and better computatio
447 followers
1,139 followers
RT @arxiv_cs_LG: Linear Self-Attention Approximation via Trainable Feedforward Kernel. Uladzislau Yorsh and Alexander Kovalenko https://t.c…
1,891 followers
We aim to expand the idea oftrainable kernel methods to approximate the self-attention mechanism of the Efficient Transformers architecture. 📄 https://t.co/zVd9IDVrYF https://t.co/rnHtnO4nTO
1,163 followers
"Linear Self-Attention Approximation via Trainable Feedforward Kernel", Uladzislau Yorsh, Alexander Kovalenko https://t.co/hzF7yVnrpi