Crossmark

Mitigating catastrophic forgetting in lifelong learning: a hybrid architecture integrating neural ordinary differential equations with memory-augmented transformers

Crossref DOI link: https://doi.org/10.1038/s41598-025-31685-9

Published Online: 2025-12-15

Update policy: https://doi.org/10.1007/springer_crossmark_policy

Authors

Zhou, Song

Li, Qiang
License Information

Text and Data Mining valid from 2025-12-15

Version of Record valid from 2026-01-15
More Information

Article History

Received: 12 October 2025

Accepted: 4 December 2025

First Online: 15 December 2025

Competing interests

: The authors declare no competing interests.

: This research involves computational experiments on publicly available benchmark datasets and does not involve human subjects, animal subjects, or clinical trials. The study utilizes standard machine learning datasets (Split CIFAR-100, Permuted MNIST, and CORe50) that are publicly accessible and widely used in the research community. All experiments were conducted in accordance with institutional guidelines for computational research and data usage policies. The research protocol was reviewed and approved by the Institutional Review Board of Jiaxing Vocational & Technical College (Approval Number: JVTC-CS-2024-087, Approval Date: March 15, 2024). No personally identifiable information or sensitive data was collected or processed during this study.

Document is current

Any future updates will be listed below