Home

Published

- 2 min read

Task-Incremental Learning on Long Text Sequences

NeuralNetworks NaturalLanguage
img of Task-Incremental Learning on Long Text Sequences

Conference Paper

Details

Can we continually learn from a stream of language-oriented tasks using Large Language Models? What is the input sequences are somewhat longer than usual? What is the impact of Task Arithmetic in LoRA-based models?

These are the challenging research questions addressed in this paper!

  • Authors: Natalia Graziuso, Andrea Zugarini, Stefano Melacci
  • Title: Task-Incremental Learning on Long Text Sequences
  • Where: Italian Conference on Computational Linguistics (CLiC-it) 2024

BibTeX

   @inproceedings{cmoss,
  author       = {Natalia Graziuso and Andrea Zugarini and Stefano Melacci},
  editor       = {Felice Dell'Orletta and Alessandro Lenci and Simonetta Montemagni and Rachele Sprugnoli},
  title        = {Task-Incremental Learning on Long Text Sequences},
  booktitle    = {{Proceedings of the Tenth Italian Conference on Computational Linguistics (CLiC-it 2024)}},
  series       = {CEUR Workshop Proceedings (CEUR-WS.org, ISSN 1613-0073)},
  volume       = {TBA},
  pages        = {TBA},
  publisher    = {CEUR},
  year         = {2024},
  url          = {https://clic2024.ilc.cnr.it/wp-content/uploads/2024/12/48_main_long.pdf},
  doi          = {TBA}
}