iMTE

논문 제목 : Grad-CAM: Why did you say that? 논문 주소 : https://arxiv.org/abs/1611.07450 Grad-CAM: Why did you say that? We propose a technique for making Convolutional Neural Network (CNN)-based models more transparent by visualizing input regions that are 'important' for predictions -- or visual explanations. Our approach, called Gradient-weighted Class Activation Mapping arxiv.org 주요 내용 정리: 1) Grad-C..

논문 제목 : Grad-CAM Guided Channel-spatial Attention Module for Fine-grained Visual Classification 논문 주소 : https://arxiv.org/abs/2101.09666 Grad-CAM guided channel-spatial attention module for fine-grained visual classification Fine-grained visual classification (FGVC) is becoming an important research field, due to its wide applications and the rapid development of computer vision technologies. Th..

논문 제목 : Ablation-CAM: Visual Explanations for Deep Convolutional Network Via Gradient-free Localization 논문 주소 : https://openaccess.thecvf.com/content_WACV_2020/html/Desai_Ablation-CAM_Visual_Explanations_for_Deep_Convolutional_Network_via_Gradient-free_Localization_WACV_2020_paper.html WACV 2020 Open Access Repository Ablation-CAM: Visual Explanations for Deep Convolutional Network via Gradient-..

논문 제목 : Group-CAM: Group Score-Weighted Visual Explanations for Deep Convolutional Networks 논문 주소 : https://arxiv.org/abs/2103.13859 Group-CAM: Group Score-Weighted Visual Explanations for Deep Convolutional Networks In this paper, we propose an efficient saliency map generation method, called Group score-weighted Class Activation Mapping (Group-CAM), which adopts the "split-transform-merge" str..

논문 제목 : SS-CAM: Smoothed Score-CAM for Sharper Visual Feature Localization 논문 주소 : https://arxiv.org/abs/2006.14255 SS-CAM: Smoothed Score-CAM for Sharper Visual Feature Localization Interpretation of the underlying mechanisms of Deep Convolutional Neural Networks has become an important aspect of research in the field of deep learning due to their applications in high-risk environments. To expl..

논문 제목 : Score-CAM : Score-weighted visual explanations for convolutional neural networks 논문 주소 : https://openaccess.thecvf.com/content_CVPRW_2020/html/w1/Wang_Score-CAM_Score-Weighted_Visual_Explanations_for_Convolutional_Neural_Networks_CVPRW_2020_paper.html CVPR 2020 Open Access Repository Haofan Wang, Zifan Wang, Mengnan Du, Fan Yang, Zijian Zhang, Sirui Ding, Piotr Mardziel, Xia Hu; Proceedi..

논문 제목 : Adapting Grad-CAM for Embedding Networks 논문 주소 : https://openaccess.thecvf.com/content_WACV_2020/html/Chen_Adapting_Grad-CAM_for_Embedding_Networks_WACV_2020_paper.html WACV 2020 Open Access Repository Lei Chen, Jianhui Chen, Hossein Hajimirsadeghi, Greg Mori; Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2020, pp. 2794-2803 The gradient-weighte..

논문 제목 : Interpretable and fine-grained visual explanations for CNNs 논문 주소 : openaccess.thecvf.com/content_CVPR_2019/html/Wagner_Interpretable_and_Fine-Grained_Visual_Explanations_for_Convolutional_Neural_Networks_CVPR_2019_paper.html CVPR 2019 Open Access Repository Interpretable and Fine-Grained Visual Explanations for Convolutional Neural Networks Jorg Wagner, Jan Mathias Kohler, Tobias Gindel..

논문 제목 : Grad-CAM++: Generalized Gradient-based Visual Explanations for Deep Convolutional Networks 논문 주소 : arxiv.org/pdf/1710.11063.pdf IEEE WACV (2018, ieeexplore.ieee.org/document/8354201)에 나온 논문을 바탕으로 이해하고 내용을 작성한다. arixv에서 나온 버전이 좀 더 extended version임으로 Grad-CAM++에 더 깊은 이해를 위해서는 extended version을 읽는 것을 추천한다. 주요 내용 : 1) Deep models은 "black box"로서 internal function을 이해하는데에는 어려움이 있다. 이를 해결하기 위해..

Meta-learning with Implicit Gradients [1] https://papers.nips.cc/paper/2019/hash/072b030ba126b2f4b2374f342be9ed44-Abstract.html IntroductionMeta-learning의 frame에서 bi-level optimization procedure는 다음으로 나누어진다.1) inner optimization : 주어진 task에 base learner가 학습하는 과정2) outer optimization : 여러 tasks 들에서 meta learner가 학습하는 과정MAML, DAML, Reptile 등의 방법이 optimization-based methods에 속한다. (Hands-on one-shot..