MAGIC-SSCIL: Manifold Anchoring and Geometric Incremental Calibration for Semi-Supervised Class Incremental Learning

作者：Yousef Abdi, Mohammad Asadpour, Yousef Seyfari · 单位：forget catastrophically due to feature drift, and their pseudo-labels become increasingly unreliable, as the label space grows. In this paper, we propose MAGIC (Manifold Anchoring and Geometric, MAGIC's design centers on two components. The first is Soft-Weighted Geometry Calibration, (SWGC), which uses graph-based label propagation on the learner's plastic feature space to weight, plastic adapter. Across CIFAR-100, CUB-200, and ImageNet-R, at label ratios of 1%, 5%, and, low-label setting, where confidence thresholding fails most clearly., supervised learning, Label propagation, a Faculty of Electrical and Computer Engineering, Universityof Tabriz, Tabriz, Iran, c Faculty of Engineering, University of Maragheh, Maragheh,Iran, Learning (Task-IL), a task identifier is available at inference, which allows task-specific, while the label set stays fixed. The most demanding and realistic setting is Class-Incremental, defined labels, new product types, or previously unseen species in ecological monitoring. It is also · 会议/期刊：arXiv preprint · 方向：cs.CV · 发布日期：2026-08-11