Search Results for author:
Found 2 papers, 1 papers with code
Improving Pre-trained Language Model Sensitivity via Mask Specific losses: A case study on Biomedical NER
To address insensitive fine-tuning, we propose Mask Specific Language Modeling (MSLM), an approach that efficiently acquires target domain knowledge by appropriately weighting the importance of domain-specific terms (DS-terms) during fine-tuning.
Clinical Prompt Learning with Frozen Language Models
Prompt learning is a new paradigm in the Natural Language Processing (NLP) field which has shown impressive performance on a number of natural language tasks with common benchmarking text datasets in full, few-shot, and zero-shot train-evaluation setups.
Cannot find the paper you are looking for? You can
Submit a new open access paper.
