Portrait de Megh Thakkar n'est pas disponible

Megh Thakkar

megh.thakkar@mila.quebec
Maîtrise recherche - UdeM
Superviseur⋅e principal⋅e
Sarath Chandar
Sujets de recherche
Traitement du langage naturel

Publications

Self-Influence Guided Data Reweighting for Language Model Pre-training
Megh Thakkar
Tolga Bolukbasi
Sriram Ganapathy
Shikhar Vashishth
Sarath Chandar
Partha Talukdar
Language Models (LMs) pre-trained with self-supervision on large text corpora have become the default starting point for developing models f… (voir plus)or various NLP tasks. Once the pre-training corpus has been assembled, all data samples in the corpus are treated with equal importance during LM pre-training. However, due to varying levels of relevance and quality of data, equal importance to all the data samples may not be the optimal choice. While data reweighting has been explored in the context of task-specific supervised learning and LM fine-tuning, model-driven reweighting for pre-training data has not been explored. We fill this important gap and propose PRESENCE, a method for jointly reweighting samples by leveraging self-influence (SI) scores as an indicator of sample importance and pre-training. PRESENCE promotes novelty and stability for model pre-training. Through extensive analysis spanning multiple model sizes, datasets, and tasks, we present PRESENCE as an important first step in the research direction of sample reweighting for pre-training language models.
2023-11-30
Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing (publié)
doi.org
openreview.net