티스토리 뷰
ProCa는 shared class detection 이라고 해서 positive negative라는 샘플연결짓기 (self supervised에서 나오는 개념있음) 를 했음
이것은 PLDCA논문에서 pseudo labeling을 통해 라벨을 classificaqtion 했던 것과는 전혀 다른 접근 법임.
근데 ProCA에서도 prototype memory bank에 넣는 과정에서 pseudo-labeling하네 ?
ProCa에서 가진 shared class detection 모듈이 PLDCA에서는 무시되었음.
왜 무시했을까? 필요가 없으니까 했겠지?
그리고 PLDCA논문에서는 ProCa가 catastrophic forgetting 이슈를 완화시켰지만 biased source knowledge로 야기된 negative transfer 이슈로
current target class를 학습하는데 심각한 영향을 끼쳤다고 한단 말임? 근데 ProCA논문에서를 그런 언급을 본 적이 없음. 이건 어떻게 안 거임?
//Supplementary Materials for
“Prototype-Guided Continual Adaptation for
Class-Incremental Unsupervised Domain
Adaptation”이것까지 봐야되려나 .
'Warble' 카테고리의 다른 글
visual studio code 주석하다 먹힐때 (1) | 2024.05.31 |
---|---|
논문 볼 때 준비물 (0) | 2024.05.15 |
층간소음때문에 (0) | 2024.04.06 |
0327 ZoDi (0) | 2024.03.28 |
한눈 팔려는 거 아니고 (2) | 2024.03.26 |
공지사항
최근에 올라온 글
최근에 달린 댓글
- Total
- Today
- Yesterday
링크
TAG
- 샴네트워크
- Group Acitivty Recognition
- /var/log/faillog
- ctfloader
- 첫논문
- alternatives.log
- Deformable Part Model
- tf2to1
- TransferLearning
- Deformable Convolutional Networks
- tensorflow convert
- Source-free unsupervised domain adaptation
- batch란
- /var/log/fontconfig.log
- 1-stage detection
- 딥페이크탐지
- StyleCLIP
- /var/log/tallylog
- yolov3 #tensorflow_convert
- torchmodel
- Arbitrary image stylization
- neural network applications
- 도메인 어댑테이션
- textual guidance
- DetectoRS
- max-margin_loss
- tf version porting
- 전이학습
- /var/log/alternatives.log
- Chatgpt4
일 | 월 | 화 | 수 | 목 | 금 | 토 |
---|---|---|---|---|---|---|
1 | 2 | 3 | 4 | 5 | ||
6 | 7 | 8 | 9 | 10 | 11 | 12 |
13 | 14 | 15 | 16 | 17 | 18 | 19 |
20 | 21 | 22 | 23 | 24 | 25 | 26 |
27 | 28 | 29 | 30 | 31 |
글 보관함