Fine-Tuning Pre-trained Language Model with Weak Supervision: A Contrastive-Regularized Self-Training Approach

21789

Knowledge Triples

Source Entity	Relationship	Target Entity
Cosine OtherEntity	appliesTo	Sentence Pair Classification Task
Cosine OtherEntity	appliesTo	Sequence Classification Task
Cosine OtherEntity	appliesTo	Token Classification Task
Fine-tuned Pre-trained Language Models Method	requires	Weak Supervision Method
Weak Supervision Method	addresses	Contrastive Self Training Method