AdaptForever: Elastic and Mutual Learning for Continuous NLP Task Mastery
Open Access
- 1 January 2025
- journal article
- Published by Tech Science Press in Computers, Materials & Continua
- Vol. 82 (3) , 4003-4019
Abstract
In natural language processing (NLP), managing multiple downstream tasks through fine-tuning pre-trained models often requires maintaining separate task-specific models, leading to practical inefficiencies. To address this challenge, we introduce AdaptForever, a novel approach that enables continuous mastery of NLP tasks through the integration of elastic and mutual learning strategies with a stochastic expert mechanism. Our method freezes the pre-trained model weights while incorporating adapters enhanced with mutual learning capabilities, facilitating effective knowledge transfer from previous tasks to new ones. By combining Elastic Weight Consolidation (EWC) for knowledge preservation with specialized regularization terms, AdaptForever successfully maintains performance on earlier tasks while acquiring new capabilities. Experimental results demonstrate that AdaptForever achieves superior performance across a continuous sequence of NLP tasks compared to existing parameter-efficient methods, while effectively preventing catastrophic forgetting and enabling positive knowledge transfer between tasks.Keywords
This publication has 16 references indexed in Scilit:
- AdaMix: Mixture-of-Adaptations for Parameter-efficient Model Tuning2022
- BitFit: Simple Parameter-efficient Fine-tuning for Transformer-based Masked Language-models2022
- Misinformation detection using multitask learning with mutual learning for novelty detection and emotion recognitionInformation Processing & Management, 2021
- Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing2021
- AdapterFusion: Non-Destructive Task Composition for Transfer Learning2021
- On the Effectiveness of Adapter-based Tuning for Pretrained Language Model Adaptation2021
- Continual Lifelong Learning in Natural Language Processing: A Survey2020
- Unsupervised Cross-lingual Representation Learning at Scale2020
- Note on the quadratic penalties in elastic weight consolidationProceedings of the National Academy of Sciences, 2018
- Overcoming catastrophic forgetting in neural networksProceedings of the National Academy of Sciences, 2017