{"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/multi-task-learning-with-group-specific","title":"Multi-Task Learning with Group-Specific Feature Space Sharing","arxiv_id":"1508.03329","date":"2015-08-13","proceeding":null,"authors":["Niloofar Yousefi","Michael Georgiopoulos","Georgios C. Anagnostopoulos"],"abstract":"When faced with learning a set of inter-related tasks from a limited amount\nof usable data, learning each task independently may lead to poor\ngeneralization performance. Multi-Task Learning (MTL) exploits the latent\nrelations between tasks and overcomes data scarcity limitations by co-learning\nall these tasks simultaneously to offer improved performance. We propose a\nnovel Multi-Task Multiple Kernel Learning framework based on Support Vector\nMachines for binary classification tasks. By considering pair-wise task\naffinity in terms of similarity between a pair's respective feature spaces, the\nnew framework, compared to other similar MTL approaches, offers a high degree\nof flexibility in determining how similar feature spaces should be, as well as\nwhich pairs of tasks should share a common feature space in order to benefit\noverall performance. The associated optimization problem is solved via a block\ncoordinate descent, which employs a consensus-form Alternating Direction Method\nof Multipliers algorithm to optimize the Multiple Kernel Learning weights and,\nhence, to determine task affinities. Empirical evaluation on seven data sets\nexhibits a statistically significant improvement of our framework's results\ncompared to the ones of several other Clustered Multi-Task Learning methods.","url_abs":"http://arxiv.org/abs/1508.03329v1","url_pdf":"http://arxiv.org/pdf/1508.03329v1.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":"multi-task-learning-with-group-specific","repo_url":"https://github.com/niloofaryousefi/ECML2015","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"binary-classification","task_name":"Binary Classification"},{"task_slug":"multi-task-learning","task_name":"Multi-Task Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}