Privacy-Preserving Models for Legal Natural Language Processing

Ying Yin, Ivan Habernal

Pre-training large transformer models with in-domain data improves domain adaptation and helps gain performance on the domain-specific downstream tasks. However, sharing models pre…