Events Calendar

PhD Thesis Proposal - Rujun Han

Name: Rujun Han


Time: 2-3:30pm, Sep 10th


Committee Members: Nanyun Peng (Chair), Aram Galstyan, Shrikanth Narayanan, Aiichiro Nakano, Xiang Ren


Title: Event-centric Reasoning in NLP with IE, QA, and NLG Applications

Abstract: Stories or narratives are composed based on events. Knowing how these events are connected logically and semantically is crucial for understanding human written texts. In the proposed thesis, we explore and invent algorithms, as well as construct novel datasets to facilitate future investigation and evaluation of automated event-centric reasoning. Specifically, we design structural deep neural networks, adapt continual pretraining methods, and leverage graph attention models to enhance machines' reasoning skills in text data. We then apply these techniques to solve event-centric tasks in information extraction, reading comprehension, question-answering, and natural language generation.

WebCast Link: https://usc.zoom.us/j/4738647635

Audiences: Everyone Is Invited

Contact: Lizsl De Leon