{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,7]],"date-time":"2026-03-07T18:25:43Z","timestamp":1772907943284,"version":"3.50.1"},"publisher-location":"California","reference-count":0,"publisher":"International Joint Conferences on Artificial Intelligence Organization","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2022,7]]},"abstract":"<jats:p>Conventional model upgrades for visual search systems require offline refresh of gallery features by feeding gallery images into new models (dubbed as \u201cbackfill\u201d), which is time-consuming and expensive, especially in large-scale applications. The task of backward-compatible representation learning is therefore introduced to support backfill-free model upgrades, where the new query features are interoperable with the old gallery features. Despite the success, previous works only investigated a close-set training scenario (i.e., the new training set shares the same classes as the old one), and are limited by more realistic and challenging open-set scenarios. To this end, we first introduce a new problem of universal backward-compatible representation learning, covering all possible data split in model upgrades. We further propose a simple yet effective method, dubbed as Universal Backward-Compatible Training (UniBCT) with a novel structural prototype refinement algorithm, to learn compatible representations in all kinds of model upgrading benchmarks in a unified manner. Comprehensive experiments on the large-scale face recognition datasets MS1Mv3 and IJB-C fully demonstrate the effectiveness of our method. Source code is available at https:\/\/github.com\/TencentARC\/OpenCompatible.<\/jats:p>","DOI":"10.24963\/ijcai.2022\/225","type":"proceedings-article","created":{"date-parts":[[2022,7,16]],"date-time":"2022-07-16T02:55:56Z","timestamp":1657940156000},"page":"1615-1621","source":"Crossref","is-referenced-by-count":12,"title":["Towards Universal Backward-Compatible Representation Learning"],"prefix":"10.24963","author":[{"given":"Binjie","family":"Zhang","sequence":"first","affiliation":[{"name":"Tsinghua University"},{"name":"ARC Lab, Tencent PCG"}]},{"given":"Yixiao","family":"Ge","sequence":"additional","affiliation":[{"name":"ARC Lab, Tencent PCG"}]},{"given":"Yantao","family":"Shen","sequence":"additional","affiliation":[{"name":"AWS\/Amazon AI"}]},{"given":"Shupeng","family":"Su","sequence":"additional","affiliation":[{"name":"ARC Lab, Tencent PCG"}]},{"given":"Fanzi","family":"Wu","sequence":"additional","affiliation":[{"name":"AWS\/Amazon AI"}]},{"given":"Chun","family":"Yuan","sequence":"additional","affiliation":[{"name":"Tsinghua University"}]},{"given":"Xuyuan","family":"Xu","sequence":"additional","affiliation":[{"name":"AI Technology Center of Tencent Video"}]},{"given":"Yexin","family":"Wang","sequence":"additional","affiliation":[{"name":"AI Technology Center of Tencent Video"}]},{"given":"Ying","family":"Shan","sequence":"additional","affiliation":[{"name":"ARC Lab, Tencent PCG"}]}],"member":"10584","event":{"name":"Thirty-First International Joint Conference on Artificial Intelligence {IJCAI-22}","theme":"Artificial Intelligence","location":"Vienna, Austria","acronym":"IJCAI-2022","number":"31","sponsor":["International Joint Conferences on Artificial Intelligence Organization (IJCAI)"],"start":{"date-parts":[[2022,7,23]]},"end":{"date-parts":[[2022,7,29]]}},"container-title":["Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence"],"original-title":[],"deposited":{"date-parts":[[2022,7,18]],"date-time":"2022-07-18T11:08:27Z","timestamp":1658142507000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.ijcai.org\/proceedings\/2022\/225"}},"subtitle":[],"proceedings-subject":"Artificial Intelligence Research Articles","short-title":[],"issued":{"date-parts":[[2022,7]]},"references-count":0,"URL":"https:\/\/doi.org\/10.24963\/ijcai.2022\/225","relation":{},"subject":[],"published":{"date-parts":[[2022,7]]}}}