Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Frustratingly Short Attention Spans in Neural Language Modeling #215

Open
icoxfog417 opened this issue Feb 17, 2017 · 1 comment
Open

Frustratingly Short Attention Spans in Neural Language Modeling #215

icoxfog417 opened this issue Feb 17, 2017 · 1 comment
Labels

Comments

@icoxfog417
Copy link
Member

icoxfog417 commented Feb 17, 2017

一言でいうと

Attentionを行う場合、隠れ層のベクトルは次の単語の予測・Attentionの算出・将来の単語に有用な情報の格納、という3つの役割を担っていることになる。なので出力を3つにして役割分担させるアイデア。併せて、単純に過去の隠れ層を結合して入力するだけでも高精度になることを確認

論文リンク

https://arxiv.org/abs/1702.04521

著者/所属機関

Michał Daniluk, Tim Rocktäschel, Johannes Welbl, Sebastian Riedel

Department of Computer Science
University College London

解説

自然言語処理における、Attentionの耐えられない短さ

@icoxfog417 icoxfog417 added the NLP label Feb 17, 2017
@icoxfog417
Copy link
Member Author

解説ブログが登場

Frustratingly Short Attention Spans (ICLR 2017) - A Summary

@icoxfog417 icoxfog417 added ICLR and removed ICLR-2017 labels Dec 8, 2017
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

1 participant