継続学習のための「睡眠」アプローチ（24 分読）

View PDF

HTML (experimental)

Abstract:The past few decades have witnessed significant advances in the design of machine learning algorithms, from early studies on task-specific shallow models to more general deep Large Language Models (LLMs). Despite showing promising results in tasks that require instant prediction or in-context learning, existing models lack the ability to continually learn and effectively transfer their temporal in-context knowledge to their long-term parameters. Inspired by human learning process, we introduce a ''Sleep'' paradigm that allows the models to continually learn, distill their short-term fragile memories into stable long-term knowledge with replay, and recursively improve themselves with ''Dreaming'' process. In more detail, sleep consists of two stages: (1) Memory Consolidation: an upward distillation process, called Knowledge Seeding, where the memories of a smaller-self are distilled into a larger network to provide more capacity while preserving the knowledge. As a proof of concept, we present a new Generalized Distillation process for {Knowledge Seeding} (i.e., the combination of on-policy distillation with Reinforcement Learning (RL)-based imitation learning); (2) Dreaming: a self-improvement phase, where the model uses RL to generate a curriculum of synthetic data to rehearse new knowledge and refine existing capabilities without human supervision. Our experiments on long-horizon, continual learning, knowledge incorporation, and few-shot generalization tasks support the importance of the sleep stage.

Comments:

A version of this work has been publicly available from September 2025 on OpenReview

Subjects:

Machine Learning (cs.LG); Artificial Intelligence (cs.AI)

Cite as:

arXiv:2606.03979 [cs.LG]

(or

arXiv:2606.03979v1 [cs.LG] for this version)

https://doi.org/10.48550/arXiv.2606.03979

arXiv-issued DOI via DataCite

Submission history

From: Ali Behrouz [view email] [v1]

Tue, 2 Jun 2026 17:56:55 UTC (2,961 KB)