HOP to the Next Tasks and Domains for Continual Learning in NLP

Authors

  • Umberto Michieli Samsung Research UK
  • Mete Ozay Samsung Research UK

DOI:

https://doi.org/10.1609/aaai.v38i13.29349

Keywords:

ML: Life-Long and Continual Learning, NLP: Text Classification, ML: Deep Neural Architectures and Foundation Models, NLP: (Large) Language Models

Abstract

Continual Learning (CL) aims to learn a sequence of problems (i.e., tasks and domains) by transferring knowledge acquired on previous problems, whilst avoiding forgetting of past ones. Different from previous approaches which focused on CL for one NLP task or domain in a specific use-case, in this paper, we address a more general CL setting to learn from a sequence of problems in a unique framework. Our method, HOP, permits to hop across tasks and domains by addressing the CL problem along three directions: (i) we employ a set of adapters to generalize a large pre-trained model to unseen problems, (ii) we compute high-order moments over the distribution of embedded representations to distinguish independent and correlated statistics across different tasks and domains, (iii) we process this enriched information with auxiliary heads specialized for each end problem. Extensive experimental campaign on 4 NLP applications, 5 benchmarks and 2 CL setups demonstrates the effectiveness of our HOP.

Published

2024-03-24

How to Cite

Michieli, U., & Ozay, M. (2024). HOP to the Next Tasks and Domains for Continual Learning in NLP. Proceedings of the AAAI Conference on Artificial Intelligence, 38(13), 14359-14369. https://doi.org/10.1609/aaai.v38i13.29349

Issue

Section

AAAI Technical Track on Machine Learning IV