Taiwei Shi

Researcher (NLP + CSS)

I teach computers to learn, think, and teach.
I have collaborated with researchers and developers while working at USC and Georgia Tech.

Education

Present — Aug. 2023
Ph.D. in Computer Science
University of Southern California, Los Angeles, CA
Advisor: Jieyu Zhao
May 2023 — Aug. 2020
B.S. in Computer Science
Georgia Institute of Technology, Atlanta, GA
Thesis: Investigating AAVE in Question Answering System
Overall GPA: 3.96/4.00, Highest Honors
Thesis
May 2020 — Aug. 2017
High School Diploma
George School, Newtown, PA
Head of School List, Honor Roll

Academic Research Experience

Present - Aug 2023
University of Southern California, Los Angeles, CA
Research Assistant, Language, Intelligence, and Model Ethics (LIME) Lab
Advisor: Jieyu Zhao
Working on efficient LLMs, responsible NLP, and computational social science.
May 2023 - Aug 2022
Georgia Institute of Technology, Atlanta, GA
Research Assistant, Entertainment Intelligence and Human-Centered AI Labs
Advisor: Mark Riedl
Worked on explainable AI and controlled story generation.
Dec 2022 — May 2022
USC Information Sciences Institute, Marina del Rey, CA
Research Intern, CUTELABNAME
Advisor: Jonathan May, Xuezhe Ma
Improved automated online content moderation.
Aug. 2022 — Aug. 2021
Georgia Institute of Technology, Atlanta, GA
Research Assistant, Social and Language Technologies (SALT) Lab
Advisor: Diyi Yang
Explored the robustness of QA and deep generative models on different dialects.
Jun. 2021 — Mar. 2021
Nanyang Technological University, Remote
Research Assistant, NTU NLP Group
Advisor: Luu Anh Tuan
Contextualized hate speech classifiers with a novel regularization technique.

Honors and Awards

2022
Georgia Tech Convergence Innovation Competition Runner-up
For my iOS keyboard extension that encourages positive thinking.
2020 — 2023
Georgia Tech Dean's List
Achieved at least a 3.5 GPA during a semester with minimum 14 credit hours.

Publications

Selected: Latest & Greatest

Can Language Model Moderators Improve the Health of Online Discourse?
Hyundong Cho, Shuai Liu, Taiwei Shi, Darpan Jain, Basem Rizk, Yuyang Huang, Zixun Lu, Nuan Wen, Jonathan Gratch, Emilio Ferrara, Jonathan May
North American Chapter of the Association for Computational Linguistics (NAACL). 2024.
Project PDF BibTeX
Safer-Instruct: Aligning Language Models with Automated Preference Data
Taiwei Shi, Kai Chen, Jieyu Zhao
North American Chapter of the Association for Computational Linguistics (NAACL). 2024.
Project PDF Code BibTeX
CoAnnotating: Uncertainty-Guided Work Allocation between Human and Large Language Models for Data Annotation
Minzhi Li, Taiwei Shi, Caleb Ziems, Min-Yen Kan, Nancy F. Chen, Zhengyuan Liu, Diyi Yang
Empirical Methods in Natural Language Processing (EMNLP). 2023.
Project PDF Code BibTeX

Conference

C3
Can Language Model Moderators Improve the Health of Online Discourse?
Hyundong Cho, Shuai Liu, Taiwei Shi, Darpan Jain, Basem Rizk, Yuyang Huang, Zixun Lu, Nuan Wen, Jonathan Gratch, Emilio Ferrara, Jonathan May
North American Chapter of the Association for Computational Linguistics (NAACL). 2024.
Project PDF BibTeX
C2
Safer-Instruct: Aligning Language Models with Automated Preference Data
Taiwei Shi, Kai Chen, Jieyu Zhao
North American Chapter of the Association for Computational Linguistics (NAACL). 2024.
Project PDF Code BibTeX
C1
CoAnnotating: Uncertainty-Guided Work Allocation between Human and Large Language Models for Data Annotation
Minzhi Li, Taiwei Shi, Caleb Ziems, Min-Yen Kan, Nancy F. Chen, Zhengyuan Liu, Diyi Yang
Empirical Methods in Natural Language Processing (EMNLP). 2023.
Project PDF Code BibTeX

Preprint

P1
How Susceptible are Large Language Models to Ideological Manipulation?
Kai Chen, Zihao He, Jun Yan, Taiwei Shi, Kristina Lerman
Preprint (Preprint). 2024.
Project PDF Code BibTeX

Workshop

W1
Neural Story Planning
Anbang Ye, Christopher Cui, Taiwei Shi, Mark Riedl
Workshop on Creative AI at Association for the Advancement of Artificial Intelligence (AAAI). 2023.
Project PDF BibTeX

Talks

Improving Online Moderation via Nonviolent Communication
Aug. 2022
USC Information Sciences Institute, NLP Seminar

References

Dr. Jonathan May, Associate Professor
CUTELABNAME
USC Information Sciences Institute
Dr. Mark Riedl, Professor
Entertainment Intelligence and Human-Centered AI Lab
Georgia Institute of Technology
Dr. Diyi Yang, Assistant Professor
Stanford NLP Group
Stanford Univeristy