{"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/deep-ensembles-for-low-data-transfer-learning-1","title":"Deep Ensembles for Low-Data Transfer Learning","arxiv_id":"2010.06866","date":"2020-10-14","proceeding":null,"authors":["Basil Mustafa","Carlos Riquelme","Joan Puigcerver","André Susano Pinto","Daniel Keysers","Neil Houlsby"],"abstract":"In the low-data regime, it is difficult to train good supervised models from scratch. Instead practitioners turn to pre-trained models, leveraging transfer learning. Ensembling is an empirically and theoretically appealing way to construct powerful predictive models, but the predominant approach of training multiple deep networks with different random initialisations collides with the need for transfer via pre-trained weights. In this work, we study different ways of creating ensembles from pre-trained models. We show that the nature of pre-training itself is a performant source of diversity, and propose a practical algorithm that efficiently identifies a subset of pre-trained models for any downstream dataset. The approach is simple: Use nearest-neighbour accuracy to rank pre-trained models, fine-tune the best ones with a small hyperparameter sweep, and greedily construct an ensemble to minimise validation cross-entropy. When evaluated together with strong baselines on 19 different downstream tasks (the Visual Task Adaptation Benchmark), this achieves state-of-the-art performance at a much lower inference budget, even when selecting from over 2,000 pre-trained models. We also assess our ensembles on ImageNet variants and show improved robustness to distribution shift.","url_abs":"https://arxiv.org/abs/2010.06866v2","url_pdf":"https://arxiv.org/pdf/2010.06866v2.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":[],"tasks":[{"task_slug":"diversity","task_name":"Diversity"},{"task_slug":"image-classification","task_name":"Image Classification"},{"task_slug":"transfer-learning","task_name":"Transfer Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/image-classification-on-vtab-1k-1","task":"Image Classification","dataset":"VTAB-1k","model":"ExpertEnsembles R101","rank_in_archive_order":5,"of":34,"metrics":{"Top-1 Accuracy":"77.6"},"uses_additional_data":true}],"syntology":{"syntology_url":"https://syntology.ai/paper/2010.06866","atlas_url":"https://app.syntology.ai/?focus=2010.06866","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}