{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,18]],"date-time":"2026-03-18T02:39:36Z","timestamp":1773801576200,"version":"3.50.1"},"reference-count":0,"publisher":"Association for the Advancement of Artificial Intelligence (AAAI)","issue":"10","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["AAAI"],"abstract":"<jats:p>Recent studies have revealed Neural Collapse (NC) in deep classifiers, where last-layer weights and features align into an equiangular tight frame (ETF), concentrating class information along  specific embedding directions. However, conventional fine-tuning typically disregards this structure, initializing task-specific classifier heads randomly. To explicitly leverage this phenomenon, we propose a simple yet effective method for metric learning: (1) initializing the classifier head along each class\u2019s NC direction from a pretrained model to preserve the emergent structure, and (2) injecting small isotropic Gaussian noise during finetuning to boost generalization. In addition, we provide a theoretical bound proving that our method explicitly reduces cumulative weight drift from the NC-initialization, compared to standard finetuning. This suggests that our method better preserves the pretrained model\u2019s class-specific structure. Empirically, this structural preservation yields Recall@K gains: reduced weight drift correlates with better performance. Concurrent decreases in the Neural Collapse 1 (NC1) measure confirm that stronger intra\u2010class cohesion underlies these improvements. Furthermore, we validate the effectiveness of our method on class\u2010imbalanced benchmarks.<\/jats:p>","DOI":"10.1609\/aaai.v40i10.37777","type":"journal-article","created":{"date-parts":[[2026,3,17]],"date-time":"2026-03-17T23:43:31Z","timestamp":1773791011000},"page":"8287-8295","source":"Crossref","is-referenced-by-count":0,"title":["Neural Collapse-Informed Initialization with Perturbation Injection in Classification-based Metric Learning"],"prefix":"10.1609","volume":"40","author":[{"given":"Jinhee","family":"Park","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Hee Bin","family":"Yoo","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Minjun","family":"Kim","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Byoung-Tak","family":"Zhang","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Junseok","family":"Kwon","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"9382","published-online":{"date-parts":[[2026,3,14]]},"container-title":["Proceedings of the AAAI Conference on Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/ojs.aaai.org\/index.php\/AAAI\/article\/download\/37777\/41739","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/ojs.aaai.org\/index.php\/AAAI\/article\/download\/37777\/41739","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,17]],"date-time":"2026-03-17T23:43:31Z","timestamp":1773791011000},"score":1,"resource":{"primary":{"URL":"https:\/\/ojs.aaai.org\/index.php\/AAAI\/article\/view\/37777"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,3,14]]},"references-count":0,"journal-issue":{"issue":"10","published-online":{"date-parts":[[2026,3,17]]}},"URL":"https:\/\/doi.org\/10.1609\/aaai.v40i10.37777","relation":{},"ISSN":["2374-3468","2159-5399"],"issn-type":[{"value":"2374-3468","type":"electronic"},{"value":"2159-5399","type":"print"}],"subject":[],"published":{"date-parts":[[2026,3,14]]}}}