article · 2023

Generative Replay Inspired by Hippocampal Memory Indexing for Continual Language Learning

Aru Maekawa, Hidetaka Kamigaito, Kotaro Funakoshi, Manabu Okumura

Cited by 4Open access

Abstract

Continual learning aims to accumulate knowledge to solve new tasks without catastrophic forgetting for previously learned tasks. Research on continual learning has led to the development of generative replay, which prevents catastrophic forgetting by generating pseudo-samples for previous tasks and learning them together with new tasks. Inspired by the biological brain, we propose the hippocampal memory indexing to enhance the generative replay by controlling sample generation using compressed features of previous training samples. It enables the generation of a specific training sample from previous tasks, thus improving the balance and quality of generated replay samples. Experimental results indicate that our method effectively controls the sample generation and consistently outperforms the performance of current generative replay methods.

Cite this paper

APA 7

Maekawa, A., Kamigaito, H., Funakoshi, K., & Okumura, M. (2023). Generative Replay Inspired by Hippocampal Memory Indexing for Continual Language Learning. https://doi.org/10.18653/v1/2023.eacl-main.65

MLA 9

Maekawa, Aru, et al. “Generative Replay Inspired by Hippocampal Memory Indexing for Continual Language Learning.” 2023, pp. 930–42, https://doi.org/10.18653/v1/2023.eacl-main.65.

Chicago

Maekawa, Aru, Hidetaka Kamigaito, Kotaro Funakoshi, and Manabu Okumura. 2023. “Generative Replay Inspired by Hippocampal Memory Indexing for Continual Language Learning.” https://doi.org/10.18653/v1/2023.eacl-main.65.

Harvard

Maekawa, A. et al. (2023) ‘Generative Replay Inspired by Hippocampal Memory Indexing for Continual Language Learning’. Available at: https://doi.org/10.18653/v1/2023.eacl-main.65.

BibTeX

@article{maekawa2023generative,
  title = {{Generative Replay Inspired by Hippocampal Memory Indexing for Continual Language Learning}},
  author = {Maekawa, Aru and Kamigaito, Hidetaka and Funakoshi, Kotaro and Okumura, Manabu},
  pages = {930--942},
  year = {2023},
  doi = {10.18653/v1/2023.eacl-main.65}
}

Made from the index record. Check names, capitalisation and page numbers against the paper. Need another style? AI for Research has over 10,000.