25 citations · 44 across the 7 of their papers we have counts for
8 papers · 1 filter
Improving Stability of Fine-Tuning Pretrained Language Models via Component-Wise Gradient Norm Clipping
Chenghao Yang, Xuezhe Ma
Fine-tuning over large pretrained language models (PLMs) has established many state-of-the-art results. Despite its superior performance, such fine-tuning can be unstable, resultin…
Weakly-Supervised Methods for Suicide Risk Assessment: Role of Related Domains
Chenghao Yang, Yudong Zhang, Smaranda Muresan
Social media has become a valuable resource for the study of suicidal ideation and the assessment of suicide risk. Among social media platforms, Reddit has emerged as the most prom…
Narrative Question Answering with Cutting-Edge Open-Domain QA Techniques: A Comprehensive Study
Xiangyang Mou, Chenghao Yang, Mo Yu +4
Recent advancements in open-domain question answering (ODQA), i.e., finding answers from large open-domain corpus like Wikipedia, have led to human-level performance on many datase…
Frustratingly Hard Evidence Retrieval for QA Over Books
Xiangyang Mou, Mo Yu, Bingsheng Yao +4
A lot of progress has been made to improve question answering (QA) in recent years, but the special problem of QA over narrative book stories has not been explored in-depth. We for…
Word-level Textual Adversarial Attacking as Combinatorial Optimization
Yuan Zang, Fanchao Qi, Chenghao Yang +4
Adversarial attacks are carried out to reveal the vulnerability of deep neural networks. Textual adversarial attacking is challenging because text is discrete and a small perturbat…
Modeling Semantic Compositionality with Sememe Knowledge
Fanchao Qi, Junjie Huang, Chenghao Yang +4
Semantic compositionality (SC) refers to the phenomenon that the meaning of a complex linguistic unit can be composed of the meanings of its constituents. Most related works focus…