This paper proposes modifications to the transformer architecture that improve its performance on long-range tasks, narrowing the gap with specialized layers. The modifications incorporate an inductive bias towards smoothness and locality, improving results without additional computation or parameters.

https://arxiv.org/abs//2311.16620

YouTube: https://www.youtube.com/@ArxivPapers

TikTok: https://www.tiktok.com/@arxiv_papers

Apple Podcasts: https://podcasts.apple.com/us/podcast/arxiv-papers/id1692476016

Spotify: https://podcasters.spotify.com/pod/show/arxiv-papers