{"type":"ScholarlyArticle","title":"Comparison Knowledge Translation for Generalizable Image Classification","authors":[{"name":"Zunlei Feng","slug":"zunlei-feng","given":"Zunlei","family":"Feng","affiliation":"Zhejiang University; Shanghai Institute for Advanced Study of Zhejiang University; Alibaba-Zhejiang University Joint Research Institute of Frontier Technologies"},{"name":"Tian Qiu","slug":"tian-qiu","given":"Tian","family":"Qiu","affiliation":"Zhejiang University"},{"name":"Sai Wu","slug":"sai-wu","given":"Sai","family":"Wu","affiliation":"Zhejiang University"},{"name":"Xiaotuan Jin","slug":"xiaotuan-jin","given":"Xiaotuan","family":"Jin","affiliation":"Hangzhou Honghua Digital Technology Co. , Ltd."},{"name":"Zengliang He","slug":"zengliang-he","given":"Zengliang","family":"He","affiliation":"Hangzhou Honghua Digital Technology Co. , Ltd."},{"name":"Mingli Song","slug":"mingli-song","given":"Mingli","family":"Song","affiliation":"Zhejiang University; Shanghai Institute for Advanced Study of Zhejiang University; Alibaba-Zhejiang University Joint Research Institute of Frontier Technologies"},{"name":"Huiqiong Wang","slug":"huiqiong-wang","given":"Huiqiong","family":"Wang","affiliation":"Ningbo Research Institute , Zhejiang University"}],"abstract":"Deep learning has recently achieved remarkable performance in image classification tasks, which depends heavily on massive annotation. However, the classification mechanism of existing deep learning models seems to contrast to humans' recognition mechanism. With only a glance at an image of the object even unknown type, humans can quickly and precisely find other same category objects from massive images, which benefits from daily recognition of various objects. In this paper, we attempt to build a generalizable framework that emulates the humans' recognition mechanism in the image classification task, hoping to improve the classification performance on unseen categories with the support of annotations of other categories. Specifically, we investigate a new task termed Comparison Knowledge Translation (CKT). Given a set of fully labeled categories, CKT aims to translate the comparison knowledge learned from the labeled categories to a set of novel categories. To this end, we put forward a Comparison Classification Translation Network (CCT-Net), which comprises a comparison classifier and a matching discriminator. The comparison classifier is devised to classify whether two images belong to the same category or not, while the matching discriminator works together in an adversarial manner to ensure whether classified results match the truth. Exhaustive experiments show that CCT-Net achieves surprising generalization ability on unseen categories and SOTA performance on target categories.","doi":"10.24963/ijcai.2022/411","url":"https://ijcai.org/proceedings/2022/411","pdf":"https://ijcai.org/proceedings/2022/0411.pdf","markdown":"https://ijcai.org/proceedings/2022/411.md","bibtex":"https://ijcai.org/proceedings/2022/bibtex/411","volume":{"slug":"2022","title":"Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2022","url":"https://ijcai.org/proceedings/2022/","year":2022,"location":"Vienna","dates":"23-29 July 2022","editors":"Luc De Raedt","isbn":"978-1-956792-00-3","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2022"},"section":"Machine Learning","track":"Main Track","number":411,"pages":{"first":2966,"last":2972},"keywords":["Machine Learning: Unsupervised Learning","Computer Vision: Transfer, low-shot, semi- and un- supervised learning","Machine Learning: Classification","Machine Learning: Semi-Supervised Learning"],"videos":[{"label":"Video #1","url":"https://slideslive.com/38984772","length_sec":352,"provider":"slideslive"},{"label":"Video #2","url":"https://slideslive.com/38986629","length_sec":392,"provider":"slideslive"}],"references":[{"position":1,"raw":"[Adam et al., 2011] C. Adam, L. Honglak, and Y. Ng An- drew. An analysis of single layer networks in unsupervised feature learning. AISTATS, 2011.","title":"Andrew. An analysis of single layer networks in unsupervised feature learning","authors":"C Adam, L Honglak, Y Ng","year":"2011","venue":"AISTATS","doi":null,"url":null,"internal_url":null},{"position":2,"raw":"[Bateni et al., 2020] P. Bateni, R. Goyal, V. Masrani, F. Wood, and L. Sigal. Improved few-shot visual classi- fication. CVPR, 2020.","title":"Improved few-shot visual classification","authors":"P Bateni, R Goyal, V Masrani, F Wood, L Sigal","year":"2020","venue":"CVPR","doi":null,"url":null,"internal_url":null},{"position":3,"raw":"[Chen et al., 2020] T. Chen, S. Kornblith, M. Norouzi, and G. Hinton. A simple framework for contrastive learning of visual representations. ICML, 2020.","title":"A simple framework for contrastive learning of visual representations","authors":"T Chen, S Kornblith, M Norouzi, G Hinton","year":"2020","venue":null,"doi":null,"url":null,"internal_url":null},{"position":4,"raw":"[Cheng et al., 2021] L. Cheng, Z. Feng, X. Wang, Y. Liu, J. Lei, and M. Song. Boundary knowledge translation based reference semantic segmentation. In IJCAI, 2021.","title":"Boundary knowledge translation based reference semantic segmentation","authors":"L Cheng, Z Feng, X Wang, Y Liu, J Lei, M Song","year":"2021","venue":"IJCAI","doi":null,"url":null,"internal_url":"/proceedings/2021/87"},{"position":5,"raw":"[David et al., 2019] Berthelot David, Carlini Nicholas, Goodfellow Ian, Papernot Nicolas, Oliver Avital, and Raffel Colin. Mixmatch: A holistic approach to semi- supervised learning. NeurIPS, 2019.","title":"Mixmatch: A holistic approach to semisupervised learning","authors":"Berthelot David, Carlini Nicholas, Goodfellow Ian, Papernot Nicolas, Oliver Avital, Raffel Colin","year":"2019","venue":null,"doi":null,"url":null,"internal_url":null},{"position":6,"raw":"[Dosovitskiy et al., 2021] A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, and N. Houlsby. An image is worth 16x16 words: Transformers for image recognition at scale. ICLR, 2021.","title":"An image is worth 16x16 words: Transformers for image recognition at scale","authors":"A Dosovitskiy, L Beyer, A Kolesnikov, D Weissenborn, N Houlsby","year":"2021","venue":"ICLR","doi":null,"url":null,"internal_url":null},{"position":7,"raw":"[Feng et al., 2021] Z. Feng, L. Cheng, X. Wang, X. Wang, Y. Liu, X. Du, and M. Song. Visual boundary knowledge translation for foreground segmentation. AAAI, 2021.","title":"Visual boundary knowledge translation for foreground segmentation","authors":"Z Feng, L Cheng, X Wang, X Wang, Y Liu, X Du, M Song","year":"2021","venue":null,"doi":null,"url":null,"internal_url":null},{"position":8,"raw":"[Gansbeke et al., 2020] W. V. Gansbeke, S. Vandenhende, S. Georgoulis, M. Proesmans, and L. V. Gool. Scan: Learning to classify images without labels. ECCV, 2020.","title":"Scan: Learning to classify images without labels","authors":"W V Gansbeke, S Vandenhende, S Georgoulis, M Proesmans, L V Gool","year":"2020","venue":"ECCV","doi":null,"url":null,"internal_url":null},{"position":9,"raw":"[Goodfellow et al., 2014] I. J. Goodfellow, J. Pouget- Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, and Bengio Y. Generative adversarial nets. NeurIPS, 2014.","title":"Generative adversarial nets","authors":"Goodfellow","year":"2014","venue":null,"doi":null,"url":null,"internal_url":null},{"position":10,"raw":"[Gou et al., 2020] J. Gou, B. Yu, SJ. Maybank, and D. Tao. Knowledge distillation: A survey. ICCV, 2020.","title":"Knowledge distillation: A survey","authors":"J Gou, B Yu, S J Maybank, D Tao","year":"2020","venue":"ICCV","doi":null,"url":null,"internal_url":null},{"position":11,"raw":"[He et al., 2016] K. He, X. Zhang, S. Ren, and J. Sun. Deep residual learning for image recognition. CVPR, 2016.","title":"Deep residual learning for image recognition","authors":"K He, X Zhang, S Ren, J Sun","year":"2016","venue":"CVPR","doi":null,"url":null,"internal_url":null},{"position":12,"raw":"[Huang et al., 2017] G. Huang, Z. Liu, Vdm Laurens, and K. Q. Weinberger. Densely connected convolutional net- works. CVPR, 2017.","title":"Densely connected convolutional networks","authors":"G Huang, Z Liu, Vdm Laurens, K Q Weinberger","year":"2017","venue":"CVPR","doi":null,"url":null,"internal_url":null},{"position":13,"raw":"[Jake et al., 2017] Snell Jake, Swersky Kevin, and S. Zemel Richard. Prototypical networks for few-shot learning. NeurIPS, 2017.","title":"Prototypical networks for few-shot learning","authors":"Jake Snell, Swersky Kevin, S Zemel Richard","year":"2017","venue":"NeurIPS","doi":null,"url":null,"internal_url":null},{"position":14,"raw":"[Jia et al., 2009] D. Jia, D. Wei, R. Socher, L. J. Li, L. Kai, and F. F. Li. Imagenet: A large-scale hierarchical image database. CVPR, 2009.","title":"Imagenet: A large-scale hierarchical image database","authors":"D Jia, D Wei, R Socher, L J Li, L Kai, F F Li","year":"2009","venue":"CVPR","doi":null,"url":null,"internal_url":null},{"position":15,"raw":"[Jolles et al., 2010] D. D. Jolles, M. J. Grol, Mav Buchem, Sarb Rombouts, and E. A. Crone. Practice effects in the brain: Changes in cerebral activation after working memory practice depend on task demands. Neuroimage, 52(2):658-668, 2010.","title":"Practice effects in the brain: Changes in cerebral activation after working memory practice depend on task demands","authors":"D D Jolles, M J Grol, Mav Buchem, Sarb Rombouts, E A Crone","year":"2010","venue":"Neuroimage","doi":null,"url":null,"internal_url":null},{"position":16,"raw":"[Koch et al., 2015] G. Koch, R. Zemel, and R. Salakhutdi- nov. Siamese neural networks for one-shot image recogni- tion. ICML, 2015.","title":"Siamese neural networks for one-shot image recognition","authors":"G Koch, R Zemel, R Salakhutdinov","year":"2015","venue":null,"doi":null,"url":null,"internal_url":null},{"position":17,"raw":"[Krizhevsky, 2009] A. Krizhevsky. Learning multiple layers of features from tiny images. Computer Science, 2009.","title":"Learning multiple layers of features from tiny images","authors":"A Krizhevsky","year":"2009","venue":"Computer Science","doi":null,"url":null,"internal_url":null},{"position":18,"raw":"[LeCun et al., 1998] Y. LeCun, L. Bottou, Y. Bengio, and P. Haffner. Gradient-based learning applied to document recognition. Proceedings of the IEEE, 1998.","title":"Gradient-based learning applied to document recognition","authors":"Y Lecun, L Bottou, Y Bengio, P Haffner","year":"1998","venue":"Proceedings of the IEEE","doi":null,"url":null,"internal_url":null},{"position":19,"raw":"[Lu et al., 2020] J. Lu, P. Gong, J. Ye, and C. Zhang. Learn- ing from very few samples: A survey. arXiv, 2020.","title":"Learning from very few samples: A survey","authors":"J Lu, P Gong, J Ye, C Zhang","year":"2020","venue":null,"doi":null,"url":null,"internal_url":null},{"position":20,"raw":"[Niyaz et al., 2018] Q. Niyaz, W. Sun, A. Y. Javaid, and M. Alam. A deep learning approach for network intrusion detection system. Frontiers in Pharmacology, 2018.","title":"A deep learning approach for network intrusion detection system","authors":"Q Niyaz, W Sun, A Y Javaid, M Alam","year":"2018","venue":"Frontiers in Pharmacology","doi":null,"url":null,"internal_url":null},{"position":21,"raw":"[Parkhi et al., 2012] O. M. Parkhi, A. Vedaldi, A. Zisserman, and C. V. Jawahar. Cats and dogs. CVPR, 2012.","title":"Cats and dogs","authors":"O M Parkhi, A Vedaldi, A Zisserman, C V Jawahar","year":"2012","venue":"CVPR","doi":null,"url":null,"internal_url":null},{"position":22,"raw":"[Rajesh and Manikanthan, 2017] M. Rajesh and Sv Manikanthan. Annoyed realm outlook taxonomy using twin transfer learning. IJPAM, 2017.","title":"Annoyed realm outlook taxonomy using twin transfer learning","authors":"M Rajesh, Sv Manikanthan","year":"2017","venue":"IJPAM","doi":null,"url":null,"internal_url":null},{"position":23,"raw":"[Sandler et al., 2018] M. Sandler, A. Howard, M. Zhu, A. Zhmoginov, and L. C. Chen. Mobilenetv2: Inverted residuals and linear bottlenecks. CVPR, 2018.","title":"Mobilenetv2: Inverted residuals and linear bottlenecks","authors":"M Sandler, A Howard, M Zhu, A Zhmoginov, L C Chen","year":"2018","venue":"CVPR","doi":null,"url":null,"internal_url":null},{"position":24,"raw":"[Siddhant et al., 2018] A. Siddhant, A. Goyal, and A. Met- allinou. Unsupervised transfer learning for spoken lan- guage understanding in intelligent agents. AAAI, 2018.","title":"Unsupervised transfer learning for spoken language understanding in intelligent agents","authors":"A Siddhant, A Goyal, A Metallinou","year":"2018","venue":null,"doi":null,"url":null,"internal_url":null},{"position":25,"raw":"[Simoes et al., 2018] R. S. Simoes, V. G. Maltarollo, P. R. O- liveira, and K. M. Honorio. Transfer and multi-task learn- ing in qsar modeling: Advances and challenges. Frontiers in Pharmacology, 2018.","title":"Transfer and multi-task learning in qsar modeling: Advances and challenges","authors":"R S Simoes, V G Maltarollo, P R Oliveira, K M Honorio","year":"2018","venue":"Frontiers in Pharmacology","doi":null,"url":null,"internal_url":null},{"position":26,"raw":"[Simonyan et al., 2014] K. Simonyan, A. Zisserman, and et al. Very deep convolutional networks for large-scale image recognition. Computer Science, 2014.","title":"Very deep convolutional networks for large-scale image recognition","authors":"K Simonyan, A Zisserman","year":"2014","venue":"Computer Science","doi":null,"url":null,"internal_url":null},{"position":27,"raw":"[Sohn et al., 2020] K. Sohn, D. Berthelot, C. L. Li, Z. Zhang, N. Carlini, E. D. Cubuk, A. Kurakin, H. Zhang, and C. Raffel. Fixmatch: Simplifying semi-supervised learn- ing with consistency and confidence. NeurIPS, 2020.","title":"Fixmatch: Simplifying semi-supervised learning with consistency and confidence","authors":"K Sohn, D Berthelot, C L Li, Z Zhang, N Carlini, E D Cubuk, A Kurakin, H Zhang, C Raffel","year":"2020","venue":null,"doi":null,"url":null,"internal_url":null},{"position":28,"raw":"[Sung et al., 2018] F. Sung, Y. Yang, L. Zhang, T. Xiang, P. H. Torr, and T. M. Hospedales. Learning to compare: Relation network for fewshot learning. CVPR, 2018.","title":"Learning to compare: Relation network for fewshot learning","authors":"F Sung, Y Yang, L Zhang, T Xiang, P H Torr, T M Hospedales","year":"2018","venue":"CVPR","doi":null,"url":null,"internal_url":null},{"position":29,"raw":"[Vinyals et al., 2016] O. Vinyals, C. Blundell, T. Lillicrap, K. Kavukcuoglu, and D. Wierstra. Matching networks for one shot learning. NeurIPS, 2016.","title":"Matching networks for one shot learning","authors":"O Vinyals, C Blundell, T Lillicrap, K Kavukcuoglu, D Wierstra","year":"2016","venue":"NeurIPS","doi":null,"url":null,"internal_url":null},{"position":30,"raw":"[Zhuang et al., 2020] F. Zhuang, Z. Qi, K. Duan, D. Xi, and Q. He. A comprehensive survey on transfer learning. Pro- ceedings of the IEEE, 2020.","title":"A comprehensive survey on transfer learning","authors":"F Zhuang, Z Qi, K Duan, D Xi, Q He","year":"2020","venue":"Proceedings of the IEEE","doi":null,"url":null,"internal_url":null}],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}