{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/domain-generalization-by-solving-jigsaw","title":"Domain Generalization by Solving Jigsaw Puzzles","arxiv_id":"1903.06864","date":"2019-03-16","proceeding":null,"authors":["Fabio Maria Carlucci","Antonio D'Innocente","Silvia Bucci","Barbara Caputo","Tatiana Tommasi"],"abstract":"Human adaptability relies crucially on the ability to learn and merge\nknowledge both from supervised and unsupervised learning: the parents point out\nfew important concepts, but then the children fill in the gaps on their own.\nThis is particularly effective, because supervised learning can never be\nexhaustive and thus learning autonomously allows to discover invariances and\nregularities that help to generalize. In this paper we propose to apply a\nsimilar approach to the task of object recognition across domains: our model\nlearns the semantic labels in a supervised fashion, and broadens its\nunderstanding of the data by learning from self-supervised signals how to solve\na jigsaw puzzle on the same images. This secondary task helps the network to\nlearn the concepts of spatial correlation while acting as a regularizer for the\nclassification task. Multiple experiments on the PACS, VLCS, Office-Home and\ndigits datasets confirm our intuition and show that this simple method\noutperforms previous domain generalization and adaptation solutions. An\nablation study further illustrates the inner workings of our approach.","url_abs":"http://arxiv.org/abs/1903.06864v2","url_pdf":"http://arxiv.org/pdf/1903.06864v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"domain-generalization-by-solving-jigsaw","repo_url":"https://github.com/fmcarlucci/JigenDG","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"AGPL-3.0"}},{"paper_slug":"domain-generalization-by-solving-jigsaw","repo_url":"https://github.com/Emma0118/domain-generalization","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"domain-generalization","task_name":"Domain Generalization"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"object-recognition","task_name":"Object Recognition"}],"methods":[{"method_slug":"jigsaw","method_name":"Jigsaw"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/domain-generalization-on-nico-animal","task":"Domain Generalization","dataset":"NICO Animal","model":"JiGen (Resnet-18)","rank_in_archive_order":3,"of":5,"metrics":{"Accuracy":"84.95"},"uses_additional_data":false},{"leaderboard":"/sota/domain-generalization-on-nico-vehicle","task":"Domain Generalization","dataset":"NICO Vehicle","model":"ResNet-18","rank_in_archive_order":4,"of":5,"metrics":{"Accuracy":"77.39"},"uses_additional_data":false},{"leaderboard":"/sota/domain-generalization-on-pacs-2","task":"Domain Generalization","dataset":"PACS","model":"JiGen (Resnet-18)","rank_in_archive_order":94,"of":133,"metrics":{"Average Accuracy":"80.51"},"uses_additional_data":false},{"leaderboard":"/sota/domain-generalization-on-pacs-2","task":"Domain Generalization","dataset":"PACS","model":"Deep All (Resnet-18)","rank_in_archive_order":98,"of":133,"metrics":{"Average Accuracy":"79.05"},"uses_additional_data":false},{"leaderboard":"/sota/domain-generalization-on-pacs-2","task":"Domain Generalization","dataset":"PACS","model":"JiGen (Alexnet)","rank_in_archive_order":110,"of":133,"metrics":{"Average Accuracy":"73.38"},"uses_additional_data":false},{"leaderboard":"/sota/domain-generalization-on-pacs-2","task":"Domain Generalization","dataset":"PACS","model":"Deep All (Alexnet)","rank_in_archive_order":117,"of":133,"metrics":{"Average Accuracy":"71.52"},"uses_additional_data":false},{"leaderboard":"/sota/image-classification-on-colored-mnist-with","task":"Image Classification","dataset":"Colored-MNIST(with spurious correlation)","model":"JiGen","rank_in_archive_order":6,"of":6,"metrics":{"Accuracy ":"11.91"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1903.06864","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}