no code implementations • Findings (ACL) 2022 • Juanyan Wang, Manali Sharma, Mustafa Bilgic
We propose a novel approach that jointly utilizes the labels and elicited rationales for text classification to speed up the training of deep learning models with limited training data.