{"type":"ScholarlyArticle","title":"Few-Shot Learning with Part Discovery and Augmentation from Unlabeled Images","authors":[{"name":"Wentao Chen","slug":"wentao-chen","given":"Wentao","family":"Chen","affiliation":""},{"name":"Chenyang Si","slug":"chenyang-si","given":"Chenyang","family":"Si","affiliation":""},{"name":"Wei Wang","slug":"wei-wang","given":"Wei","family":"Wang","affiliation":""},{"name":"Liang Wang","slug":"liang-wang","given":"Liang","family":"Wang","affiliation":""},{"name":"Zilei Wang","slug":"zilei-wang","given":"Zilei","family":"Wang","affiliation":""},{"name":"Tieniu Tan","slug":"tieniu-tan","given":"Tieniu","family":"Tan","affiliation":""}],"abstract":"Few-shot learning is a challenging task since only few instances are given for recognizing an unseen class. One way to alleviate this problem is to acquire a strong inductive bias via meta-learning on similar tasks. In this paper, we show that such inductive bias can be learned from a flat collection of unlabeled images, and instantiated as transferable representations among seen and unseen classes.   Specifically, we propose a novel part-based self-supervised representation learning scheme to learn transferable representations by maximizing the similarity of an image to its discriminative part. To mitigate the overfitting in few-shot classification caused by data scarcity, we further propose a part augmentation strategy by retrieving extra images from a base dataset. We conduct systematic studies on miniImageNet and tieredImageNet benchmarks. Remarkably, our method yields impressive results, outperforming the previous best unsupervised methods by 7.74% and 9.24% under 5-way 1-shot and 5-way 5-shot settings, which are comparable with state-of-the-art supervised methods.","doi":"10.24963/ijcai.2021/313","url":"https://ijcai.org/proceedings/2021/313","pdf":"https://ijcai.org/proceedings/2021/0313.pdf","markdown":"https://ijcai.org/proceedings/2021/313.md","bibtex":"https://ijcai.org/proceedings/2021/bibtex/313","volume":{"slug":"2021","title":"Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2021","url":"https://ijcai.org/proceedings/2021/","year":2021,"location":"Montreal","dates":"19-27 August 2021","editors":"Prof. Zhi-Hua Zhou","isbn":"978-0-9992411-9-6","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2021"},"section":"Machine Learning","track":"Main Track","number":313,"pages":{"first":2271,"last":2277},"keywords":["Machine Learning: Classification","Machine Learning: Deep Learning","Machine Learning: Unsupervised Learning"],"videos":[],"references":[],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}