Online Task-Free Continual Learning via Dynamic Expansionable Memory Distribution

TY - GEN

T1 - Online Task-Free Continual Learning via Dynamic Expansionable Memory Distribution

AU - Ye, Fei

AU - Bors, Adrian Gheorghe

N1 - This is an author-produced version of the published paper. Uploaded in accordance with the University’s Research Publications and Open Access policy.

PY - 2025/6/14

Y1 - 2025/6/14

N2 - Recent continuous learning (CL) research primarily addresses catastrophic forgetting within a straightforward learning framework where class and task information are predefined. However, in the Task-Free Continual Learning (TFCL), representing a more realistic and challenging CL scenarios, such information is typically absent. In this paper , we address the online TFCL by introducing an innovative memory management approach, by incorporating a dynamic memory system for storing selected data representatives from evolving distributions while a dynamically ex-pandable memory system enables the retention of essential long-term knowledge. The proposed dynamic expandable memory system manages a series of memory distributions, each designed to represent the information from a distinct data category. A new memory expansion mechanism that assesses the proximity between incoming samples and existing memory distributions is proposed for evaluating when to add new memory distributions into the system. Additionally , a novel memory distribution augmentation technique is proposed for selectively gathering suitable samples for each memory distribution, enhancing the statistical robustness over time. To prevent memory saturation before the training phase, we introduce a memory distribution reduction strategy that automatically eliminates overlapping memory distributions, ensuring adequate capacity for accommodating new information in subsequent learning episodes. We conduct a series of experiments demonstrating that our proposed approach attains state-of-the-art performance in both supervised and unsupervised learning contexts. The source code is available at https://github.com/dtuzi123/DEMD.

AB - Recent continuous learning (CL) research primarily addresses catastrophic forgetting within a straightforward learning framework where class and task information are predefined. However, in the Task-Free Continual Learning (TFCL), representing a more realistic and challenging CL scenarios, such information is typically absent. In this paper , we address the online TFCL by introducing an innovative memory management approach, by incorporating a dynamic memory system for storing selected data representatives from evolving distributions while a dynamically ex-pandable memory system enables the retention of essential long-term knowledge. The proposed dynamic expandable memory system manages a series of memory distributions, each designed to represent the information from a distinct data category. A new memory expansion mechanism that assesses the proximity between incoming samples and existing memory distributions is proposed for evaluating when to add new memory distributions into the system. Additionally , a novel memory distribution augmentation technique is proposed for selectively gathering suitable samples for each memory distribution, enhancing the statistical robustness over time. To prevent memory saturation before the training phase, we introduce a memory distribution reduction strategy that automatically eliminates overlapping memory distributions, ensuring adequate capacity for accommodating new information in subsequent learning episodes. We conduct a series of experiments demonstrating that our proposed approach attains state-of-the-art performance in both supervised and unsupervised learning contexts. The source code is available at https://github.com/dtuzi123/DEMD.

M3 - Conference contribution

SP - 20512

EP - 20522

BT - IEEE/CVP Conference on Computer Vision and Pattern Recognition (CVPR)

PB - IEEE

CY - Nashville, TN, USA

ER -