eprintid: 10151383 rev_number: 13 eprint_status: archive userid: 699 dir: disk0/10/15/13/83 datestamp: 2022-07-07 15:25:05 lastmod: 2022-12-05 15:34:09 status_changed: 2022-07-07 15:25:05 type: proceedings_section metadata_visibility: show sword_depositor: 699 creators_name: Wang, R creators_name: Pontil, M creators_name: Ciliberto, C title: The Role of Global Labels in Few-Shot Classification and How to Infer Them ispublished: pub divisions: C05 divisions: F48 divisions: B04 divisions: UCL note: This version is the version of record. For information on re-use, please refer to the publisher’s terms and conditions. abstract: Few-shot learning is a central problem in meta-learning, where learners must quickly adapt to new tasks given limited training data. Recently, feature pre-training has become a ubiquitous component in state-of-the-art meta-learning methods and is shown to provide significant performance improvement. However, there is limited theoretical understanding of the connection between pre-training and meta-learning. Further, pre-training requires global labels shared across tasks, which may be unavailable in practice. In this paper, we show why exploiting pre-training is theoretically advantageous for meta-learning, and in particular the critical role of global labels. This motivates us to propose Meta Label Learning (MeLa), a novel meta-learning framework that automatically infers global labels to obtains robust few-shot models. Empirically, we demonstrate that MeLa is competitive with existing methods and provide extensive ablation experiments to highlight its key properties. date: 2021 date_type: published publisher: NeurIPS official_url: https://proceedings.neurips.cc/paper/2021/file/e3b6fb0fd4df098162eede3313c54a8d-Paper.pdf oa_status: green full_text_type: pub language: eng primo: open primo_central: open_green verified: verified_manual elements_id: 1882163 lyricists_name: Pontil, Massimiliano lyricists_id: MPONT27 actors_name: Jenkins, Alexander actors_name: Flynn, Bernadette actors_id: AJENK21 actors_id: BFFLY94 actors_role: owner actors_role: impersonator full_text_status: public publication: Proceedings of the 35th Conference on Neural Information Processing Systems (NeurIPS 2021) volume: 32 pagerange: 27160-27170 book_title: Proceedings of the 35th Conference on Neural Information Processing Systems (NeurIPS 2021) citation: Wang, R; Pontil, M; Ciliberto, C; (2021) The Role of Global Labels in Few-Shot Classification and How to Infer Them. In: Proceedings of the 35th Conference on Neural Information Processing Systems (NeurIPS 2021). (pp. pp. 27160-27170). NeurIPS Green open access document_url: https://discovery.ucl.ac.uk/id/eprint/10151383/1/the_role_of_global_labels_in_f.pdf