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

Towards Maximizing the Representation Gap between In-Domain & Out-of-Distribution Examples #41

Open
nocotan opened this issue Feb 25, 2021 · 0 comments

Comments

@nocotan
Copy link
Member

nocotan commented Feb 25, 2021

一言でいうと

in-domainサンプルとOODサンプルの表現ギャップを最大化するDPNモデルのための新しい損失関数を提案.

論文リンク

https://papers.nips.cc/paper/2020/file/68d3743587f71fbaa5062152985aff40-Paper.pdf

著者/所属機関

Jay Nandy et al.
(National University of Singapore)

投稿日付(yyyy/MM/dd)

2020/12

概要

既存のDPNモデルの定式化では,複数のクラス間でデータの不確実性が高いin-domainサンプルと,OODサンプルとの間で区別のつかない表現になることが多い.
本研究では、in-domainサンプルとOODサンプルの表現ギャップを最大化するDPNモデルのための新しい損失関数を提案.

Screen Shot 2021-02-26 at 3 18 50

新規性・差分

  • 予測の不確実性の原因を特定することで情報に基づいた方法での手動介入が可能になり,実世界での応用におけるAIシステムの信頼性を高めることができる.

手法

Screen Shot 2021-02-26 at 3 18 56

Screen Shot 2021-02-26 at 3 19 01

Screen Shot 2021-02-26 at 3 19 07

結果

Screen Shot 2021-02-26 at 3 19 13

Screen Shot 2021-02-26 at 3 19 21

Screen Shot 2021-02-26 at 3 19 26

Screen Shot 2021-02-26 at 3 19 32

Screen Shot 2021-02-26 at 3 19 36

コメント

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

1 participant