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

Universal Domain Adaptation through Self-Supervision #28

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

Universal Domain Adaptation through Self-Supervision #28

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

Comments

@nocotan
Copy link
Member

nocotan commented Feb 25, 2021

一言でいうと

論文リンク

https://papers.nips.cc/paper/2020/file/bb7946e7d85c81a9e69fee1cea4a087c-Paper.pdf

著者/所属機関

Kuniaki Saito et al.
(Boston University)

投稿日付(yyyy/MM/dd)

2020/12

概要

教師なしドメイン適応の手法は典型的には全てのソースカテゴリがターゲットドメインに出現することを仮定している.
実際には,二つのドメインのオーバーラップは小さい場合も多い.
論文では任意のカテゴリシフトを扱うことができるd Domain Adaptative Neighborhood Clustering via Entropy optimization (DANCE)を提案.

Screen Shot 2021-02-26 at 1 10 59

新規性・差分

提案手法は

  •  ソースカテゴリに完全には依存せず,自己学習に基づいて近傍クラスタリングによってターゲットドメインの構造を学習
  • エントロピーに基づく特徴空間を近づける

手法

Screen Shot 2021-02-26 at 1 11 11

結果

Screen Shot 2021-02-26 at 1 11 19

Screen Shot 2021-02-26 at 1 11 27

Screen Shot 2021-02-26 at 1 11 37

Screen Shot 2021-02-26 at 1 11 46

コメント

@nocotan nocotan self-assigned this Feb 25, 2021
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