Saved in:
Bibliographic Details
Main Authors: Yanowsky, Danit, Weinshall, Daphna
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2604.08336
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910116808228864
author Yanowsky, Danit
Weinshall, Daphna
author_facet Yanowsky, Danit
Weinshall, Daphna
contents Catastrophic forgetting remains a key challenge in Continual Learning (CL). In replay-based CL with severe memory constraints, performance critically depends on the sample selection strategy for the replay buffer. Most existing approaches construct memory buffers using embeddings learned under supervised objectives. However, class-agnostic, self-supervised representations often encode rich, class-relevant semantics that are overlooked. We propose a new method, Multiple Embedding Replay Selection, MERS, which replaces the buffer selection module with a graph-based approach that integrates both supervised and self-supervised embeddings. Empirical results show consistent improvements over SOTA selection strategies across a range of continual learning algorithms, with particularly strong gains in low-memory regimes. On CIFAR-100 and TinyImageNet, MERS outperforms single-embedding baselines without adding model parameters or increasing replay volume, making it a practical, drop-in enhancement for replay-based continual learning.
format Preprint
id arxiv_https___arxiv_org_abs_2604_08336
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Leveraging Complementary Embeddings for Replay Selection in Continual Learning with Small Buffers
Yanowsky, Danit
Weinshall, Daphna
Machine Learning
Catastrophic forgetting remains a key challenge in Continual Learning (CL). In replay-based CL with severe memory constraints, performance critically depends on the sample selection strategy for the replay buffer. Most existing approaches construct memory buffers using embeddings learned under supervised objectives. However, class-agnostic, self-supervised representations often encode rich, class-relevant semantics that are overlooked. We propose a new method, Multiple Embedding Replay Selection, MERS, which replaces the buffer selection module with a graph-based approach that integrates both supervised and self-supervised embeddings. Empirical results show consistent improvements over SOTA selection strategies across a range of continual learning algorithms, with particularly strong gains in low-memory regimes. On CIFAR-100 and TinyImageNet, MERS outperforms single-embedding baselines without adding model parameters or increasing replay volume, making it a practical, drop-in enhancement for replay-based continual learning.
title Leveraging Complementary Embeddings for Replay Selection in Continual Learning with Small Buffers
topic Machine Learning
url https://arxiv.org/abs/2604.08336