EUVIMEDEUVIMED entdecken
Crossref

Revisiting Class-Incremental Learning with Pre-Trained Models: Generalizability and Adaptivity are All You Need

Da-Wei Zhou, Zi-Wen Cai, Han-Jia Ye, De-Chuan Zhan, Ziwei Liu

International Journal of Computer Vision · 2025-03 · 133 (3) · 1012-1032

DOI: 10.1007/s11263-024-02218-0 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Computer Vision, online erschienen am 31.08.2024. Verfasst von Da-Wei Zhou und 4 weiteren Autor:innen. Die Arbeit wurde bisher 179-mal zitiert (Crossref) und nennt 129 Literaturangaben. Gefördert von: National Science and Technology Major Project, NSFC, Collaborative Innovation Center of Novel Software Technology and Industrialization.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Class-incremental learning (CIL) aims to adapt to emerging new classes without forgetting old ones. Traditional CIL models are trained from scratch to continually acquire knowledge as data evolves. Recently, pre-training has achieved substantial progress, making vast pre-trained models (PTMs) accessible for CIL. Contrary to traditional methods, PTMs possess generalizable embeddings, which can be easily transferred for CIL. In this work, we revisit CIL with PTMs and argue that the core factors in CIL are adaptivity for model updating and generalizability for knowledge transferring. (1) We first reveal that frozen PTM can already provide generalizable embeddings for CIL. Surprisingly, a simple baseline (SimpleCIL) which continually sets the classifiers of PTM to prototype features can beat state-of-the-art even without training on the downstream task. (2) Due to the distribution gap between pre-trained and downstream datasets, PTM can be further cultivated with adaptivity via model adaptation. We propose AdaPt and mERge (Aper), which aggregates the embeddings of PTM and adapted models for classifier construction. Aper is a general framework that can be orthogonally combined with any parameter-efficient tuning method, which holds the advantages of PTM’s generalizability and adapted model’s adaptivity. (3) Additionally, considering previous ImageNet-based benchmarks are unsuitable in the era of PTM due to data overlapping, we propose four new benchmarks for assessment, namely ImageNet-A, ObjectNet, OmniBenchmark, and VTAB. Extensive experiments validate the effectiveness of Aper with a unified and concise framework. Code is available at https://github.com/zhoudw-zdw/RevisitingCIL.

Förderung

  • National Science and Technology Major Project
  • NSFC
  • Collaborative Innovation Center of Novel Software Technology and Industrialization
  • Ministry of Education, Singapore, under its MOE AcRF Tier 2
  • NTU NAP
  • RIE2020 Industry Alignment Fund - Industry Collaboration Projects (IAF-ICP) Funding Initiative
  • China Scholarship Council

Zitieren

APA

Zhou, D., Cai, Z., Ye, H., Zhan, D., & Liu, Z. (2025). Revisiting Class-Incremental Learning with Pre-Trained Models: Generalizability and Adaptivity are All You Need. International Journal of Computer Vision, 133(3), 1012-1032. https://doi.org/10.1007/s11263-024-02218-0

Vancouver

Zhou D, Cai Z, Ye H, Zhan D, Liu Z. Revisiting Class-Incremental Learning with Pre-Trained Models: Generalizability and Adaptivity are All You Need. International Journal of Computer Vision. 2025;133(3):1012-1032. doi:10.1007/s11263-024-02218-0

BibTeX
@article{Zhou2025,
  author = {Zhou, Da-Wei and Cai, Zi-Wen and Ye, Han-Jia and Zhan, De-Chuan and Liu, Ziwei},
  title = {Revisiting Class-Incremental Learning with Pre-Trained Models: Generalizability and Adaptivity are All You Need},
  journal = {International Journal of Computer Vision},
  year = {2025},
  volume = {133},
  number = {3},
  pages = {1012--1032},
  doi = {10.1007/s11263-024-02218-0}
}
Kostenlos starten →