Abstract:Class-incremental learning (CIL) with pre-trained models (PTMs) faces a critical trade-off between efficient adaptation and long-term stability. While analytic learning enables rapid, recursive closed-form updates, its efficacy is often compromised by accumulated errors and feature incompatibility. In this paper, we first conduct a systematic study to dissect the failure modes of PTM-based analytic CIL, identifying representation rigidity as the primary bottleneck. Motivated by this insight, we propose VILA, a novel dual-branch framework that advances analytic CIL via a two-level vision-language calibration strategy. Specifically, we coherently fuse plastic, task-adapted features with a frozen, universal visual anchor at the feature level through geometric calibration, and leverage cross-modal semantic priors at the decision level to rectify prediction bias. This confluence maintains analytic-learning's extreme efficiency while overcoming its inherent brittleness. Extensive experiments across eight benchmarks demonstrate that VILA consistently yields superior performance, particularly in fine-grained and long-sequence scenarios. Our framework harmonizes high-fidelity prediction with the simplicity of analytic learning. Our code is available at this https URL.
| Comments: | 20 pages, 11 figures, 9 tables. Accepted by ICML2026 |
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2602.13670 [cs.LG] |
| (or arXiv:2602.13670v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2602.13670 arXiv-issued DOI via DataCite |
From: Binyu Zhao [view email]
[v1]
Sat, 14 Feb 2026 08:32:51 UTC (1,137 KB)
[v2]
Wed, 6 May 2026 08:21:11 UTC (1,151 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。