Learning Adaptive and Expandable Mixture Model for Continual Learning
Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (33) · 27773-27781
DOI: 10.1609/aaai.v40i33.39999 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Fei Ye und 6 weiteren Autor:innen.
Abstract (Originalsprache)
Continuous learning constitutes a fundamental capability of artificial intelligence systems, enabling them to incrementally assimilate novel information without succumbing to catastrophic forgetting. Recent research has leveraged Pre-Trained Models (PTMs) to enhance continual learning efficacy. Nevertheless, prevailing methodologies typically depend on a singular pre-trained backbone and freeze all pre-trained parameters to mitigate network forgetting, thereby constraining adaptability to emerging tasks. In this study, we introduce an innovative PTM-based framework featuring a Dual-Representation Backbone Architecture (DRBA), which integrates both invariant and evolved representation networks to concurrently capture static and dynamic features. Building upon DRBA, we propose an Adaptive and Expandable Mixture Model (AEMM) that incrementally incorporates new expert modules with minimal parameter overhead to accommodate the learning of each novel task. To further augment adaptability, we develop a Dynamic Adaptive Representation Fusion Mechanism (DARFM) that processes outputs from both representation networks and autonomously generates data-driven adaptive weights, optimizing the contribution of each representation. This mechanism yields an adaptive, semantically enriched composite representation, thereby maximizing positive knowledge transfer. Additionally, we propose a Dynamic Knowledge Calibration Mechanism (DKCM), comprising prediction and representation calibration processes, to ensure consistency in both predictions and feature representations. This approach achieves a balance between stability and plasticity, even when learning complex datasets. Empirical evaluations substantiate that the proposed approach attains state-of-the-art performance.
Zitieren
APA
Ye, F., Zhong, Y., Liu, Q., Bors, A. G., Sun, J., Guo, J., & Zhou, S. (2026). Learning Adaptive and Expandable Mixture Model for Continual Learning. Proceedings of the AAAI Conference on Artificial Intelligence, 40(33), 27773-27781. https://doi.org/10.1609/aaai.v40i33.39999
Vancouver
Ye F, Zhong Y, Liu Q, Bors AG, Sun J, Guo J, et al. Learning Adaptive and Expandable Mixture Model for Continual Learning. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(33):27773-27781. doi:10.1609/aaai.v40i33.39999
BibTeX
@article{Ye2026,
author = {Ye, Fei and Zhong, YongCheng and Liu, Qihe and Bors, Adrian G. and Sun, JingLing and Guo, Jinyu and Zhou, ShiJie},
title = {Learning Adaptive and Expandable Mixture Model for Continual Learning},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {33},
pages = {27773--27781},
doi = {10.1609/aaai.v40i33.39999}
}
EUVIMED