Elastic Feature Consolidation For Cold Start Exemplar-Free Incremental Learning

Simone Magistri, Tomaso Trinci, Albin Soutif, Joost van de Weijer, Andrew D. Bagdanov

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Keywords: Computer vision, continual learning, class-incremental learning, exemplar free, lifelong learning

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.

TL;DR: We define an Empirical Feature Matrix used as a pseudo-metric we use to both regularize feature drift and update class prototypes during exemplar-free class-incremental learning.

Abstract: Exemplar-Free Class Incremental Learning (EFCIL) aims to learn from a sequence of tasks without having access to previous task data. In this paper, we consider the challenging Cold Start scenario in which insufficient data is available in the first task to learn a high-quality backbone. This is especially challenging for EFCIL since it requires high plasticity, which results in feature drift which is difficult to compensate for in the exemplar-free setting. To address this problem, we propose a simple and effective approach that consolidates feature representations by regularizing drift in directions highly relevant to previous tasks and employs prototypes to reduce task-recency bias. Our method, called Elastic Feature Consolidation (EFC), exploits a tractable second-order approximation of feature drift based on an Empirical Feature Matrix (EFM). The EFM induces a pseudo-metric in feature space which we use to regularize feature drift in important directions and to update Gaussian prototypes used in a novel asymmetric cross entropy loss which effectively balances prototype rehearsal with data from new tasks. Experimental results on CIFAR-100, Tiny-ImageNet, ImageNet-Subset and ImageNet-1K demonstrate that Elastic Feature Consolidation is better able to learn new tasks by maintaining model plasticity and significantly outperform the state-of-the-art.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Primary Area: transfer learning, meta learning, and lifelong learning

Submission Number: 1311