FTFT: Efficient and Robust Fine-Tuning by Transferring Training Dynamics

arXiv:2310.06588v2 Announce Type: replace-cross
Abstract: Despite the massive success of fine-tuning Pre-trained Language Models (PLMs), they remain susceptible to out-of-distribution input. Dataset cartography is a simple yet effective dual-model approach that improves the robustness of fine-tuned PLMs. It involves fine-tuning a model on the original training set (i.e. reference model), selecting a subset of important training instances based on the training dynamics, and fine-tuning again only on these selected examples (i.e. main model). However, this approach requires fine-tuning the same …

abstract arxiv cs.cl cs.lg dataset distribution dynamics fine-tuning language language models massive reference robust robustness set simple success training type