jin's blog

  • 홈
  • 태그
  • 미디어로그
  • 위치로그
  • 방명록

smoothGrad 1

[XAI] SmoothGrad: removing noise by adding noise

https://arxiv.org/pdf/1706.03825.pdf 입력이미지에 대한 class score의 gradient를 계산하는 과정에서, 입력에 noise를 주어 생성된 여러개의 saliency map을 average해서 smoothing 효과를 주는것. 이렇게 하면 오히려 주요하게 봐야하는 입력의 pixel은 더 강조가 되고 noise 같아 보이는 부분은 줄어들게 되었다고 말함 그럼 Noise 효과에 대해 생각해보자 단순하게 생각해보면 중요한 attribution을 가져야 하는부분이 계속 중요하게 나와서 그렇다고 생각할수있겠지만 다른 방면으로 생각해보면 이전에 integrated gradeint 연구에서 언급했던 axiom중에 sensitivity에서 non-zero attribution값을 가..

카테고리 없음 2022.01.09
이전
1
다음
더보기
프로필사진

jin's blog

Endure

  • 분류 전체보기 (50)
    • IT (44)
      • Paper (23)
      • Reinforcement Learni.. (0)
      • Probability (0)
      • Deep learning (6)
      • Spark (5)
      • Python (4)
      • Computer vision (4)
      • Data Structure (1)
    • 관심사 (2)
      • 낚시 (0)
      • 피아노 (2)
      • 일상 (0)

Tag

intergrated gradient, Deconvolution Network, Paper리뷰, vision transformer, TCAV, Learning Directed Exploration Strategies, Interpretability Beyond Feature Attribution:Quantitative Testing with Concept Activation Vectors, Quantifying Attention Flow in Transformers, R-CNN, smoothGrad, Regularizing Trajectory Optimization with Denoising Autoencoders, Fast R-CNN, Never Give Up, RL논문, XAI, Adversarial Examples Are Not Bugs, Axiomatic Attribution for Deep Networks, They Are Features, Concept vector, CAV,

최근글과 인기글

  • 최근글
  • 인기글

최근댓글

공지사항

페이스북 트위터 플러그인

  • Facebook
  • Twitter

Archives

Calendar

«   2025/06   »
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

방문자수Total

  • Today :
  • Yesterday :

Copyright © Kakao Corp. All rights reserved.

티스토리툴바