zihaoh@usc.edu
Zihao He
I'm a Research Scientist on the Monetization GenAI team at Meta. I obtained PhD in CS from USC, advised by Prof. Kristina Lerman. I was part of the the SEA Lab. I am interested in NLP, computational social science, and multimodal representation learning. Recently, I'm focused on the alignment & safety, and societal impacts of LLMs.
Previously, I received undergraduate degree in Communication Engineering from Beijing University of Posts and Telecommunications. I've spent time at Tsinghua University working with Shutao Xia. During my PhD, I've interned at TikTok, Amazon, and DiDi Global.
[Full CV] [Twitter] [Github] [Google Scholar] [Linkedin]
Recent News
- [02/2025] Started a position as Research Scientist at Meta!
- [02/2025] One paper on mitigating LLM hallucination is now on arXiv!
- [02/2025] Defended my thesis "Aligning Large Language Models with Human Perspectives" and finally Dr. He!
- [01/2025] One paper got accepted to NAACL 2025!
- [09/2024] Two papers got accepted to EMNLP 2024!
Education
Aug 2019 - Feb 2025
Ph.D. in Computer Science, University of Southern California, Los Angeles, CA, United States
Sep 2018 - June 2019 (Transferred)
M.S. in Computer Engineering, Tsinghua University, Beijing, China
Sep 2014 - June 2018
B.E. in Communication Engineering, Beijing University of Posts and Telecommunications, Beijing, China
Selected Publications & Preprints
-
Smoothing Out Hallucinations: Mitigating LLM Hallucination with Smoothed Knowledge Distillation
Hieu Nguyen, Zihao He, Shoumik Atul Gandre, Ujjwal Pasupulety, Sharanya Kumari Shivakumar, Kristina Lerman
Preprint, 2025 -
Improving and Assessing the Fidelity of Large Language Models Alignment to Online Communities
Minh Duc Chu, Zihao He, Rebecca Dorn, Kristina Lerman
To appear at NAACL, 2025 -
COMMUNITY-CROSS-INSTRUCT: Unsupervised Instruction Generation for Aligning Large Language Models to Online Communities
Zihao He, Rebecca Dorn, Siyi Guo, Minh Duc Chu, Kristina Lerman
EMNLP, 2024 -
How Susceptible are Large Language Models to Ideological Manipulation?
Kai Chen, Zihao He, Jun Yan, Taiwei Shi, Kristina Lerman
EMNLP, 2024 -
Whose Emotions and Moral Sentiments Do Language Models Reflect?
Zihao He, Siyi Guo, Ashwin Rao, Kristina Lerman
ACL-Findings, 2024
[Media Coverage1] [Media Coverage2] -
CPL-NoViD: Context-Aware Prompt-based Learning for Norm Violation Detection in Online Communities
Zihao He, Jonathan May, Kristina Lerman
ICWSM, 2024
[Code] -
IsamasRed: A Public Dataset Tracking Reddit Discussions on Israel-Hamas Conflict
Kai Chen, Zihao He, Keith Burghardt, Jingxin Zhang, Kristina Lerman
ICWSM, 2024
[Code] -
Don't Blame the Data, Blame the Model: Understanding Noise and Bias When Learning from Subjective Annotations
Abhishek Anand, Negar Mokhberian, Prathyusha Kumar, Anweasha Saha, Zihao He, Ashwin Rao, Fred Morstatter, Kristina Lerman
Proceedings of the 1st Workshop on Uncertainty-Aware NLP (UncertaiNLP), 2024 -
Reading Between the Tweets: Deciphering Ideological Stances of Interconnected Mixed-ideology Communities
Zihao He, Ashwin Rao, Siyi Guo, Negar Mokhberian, Kristina Lerman
EACL-Findings, 2024
[Code] -
Measuring Online Emotional Reactions to Events
Siyi Guo, Zihao He, Ashwin Rao, Eugene Jang, Yuanfeixue Nan, Fred Morstatter, Jeffrey Brantingham, Kristina Lerman
ASONAM, 2023 -
Anger Breeds Controversy: Analyzing Controversy and Emotions on Reddit
Kai Chen, Zihao He, Rong-Ching Chang, Jonathan May, Kristina Lerman
SBP-BRiMS, 2023 -
ALCAP: Alignment-Augmented Music Captioner
Zihao He, Weituo Hao, Wei-Tsung Lu, Changyou Chen, Kristina Lerman, Xuchen Song
EMNLP, 2023
[Code] -
Infusing Knowledge from Wikipedia to Enhance Stance Detection
Zihao He, Negar Mokhberian, Kristina Lerman
Proceedings of the 12th Workshop on Computational Approaches to Subjectivity, Sentiment & Social Media Analysis, 2022
[Code] -
Detecting Polarized Topics Using Partisanship-aware Contextualized Topic Embeddings
Zihao He, Negar Mokhberian, António Câmara, Andres Abeliuk, Kristina Lerman
EMNLP-Findings, 2021
[Code] [Media Coverage] -
Speaker Turn Modeling for Dialogue Act Classification
Zihao He, Leili Tavabi, Kristina Lerman, Mohammad Soleymani
EMNLP-Findings, 2021
[Code] -
Graph Embedding with Personalized Context Distribution
Zihao He, Di Huang, Yuzhong Huang, Kexuan Sun, Sami Abu-El-Haija, Bryan Perozzi, Kristina Lerman, Fred Morstatter, Aram Galstyan
Companion Proceedings of the Web Conference, 2020
[Code] -
Training interpretable convolutional neural networks by differentiating class-specific filters
Haoyu Liang, Zhihao Ouyang, Yuyuan Zeng, Hang Su, Zihao He, Shu-Tao Xia, Jun Zhu, Bo Zhang
ECCV, 2020
[Code] -
Neural Style Transfer with Content Discrimination
Xiyu Yan, Yeli Xing, Zihao He, Tao Dai, Yong Jiang, Shu-Tao Xia
ICMEW, 2019 -
Attentiondrop for Convolutional Neural Networks
Zhihao Ouyang, Yan Feng, Zihao He, Tianbo Hao, Tao Dai, Shu-Tao Xia
ICME, 2019 -
Automatic Grassland Degradation Estimation Using Deep Learning
Xiyu Yan, Yong Jiang, Shuai Chen, Zihao He, Chunmei Li, Shu-Tao Xia, Tao Dai, Shuo Dong, Feng Zheng
IJCAI, 2019
Industrial Experience
- Research Scientist Intern@TikTok, San Jose, CA, May-Aug 2023 & 2024
- Applied Scientist Intern@Amazon, Remote, May-Aug 2022
- Algorithm Development Intern@DiDi Global, Beijing, Apr-Jul 2020
Miscellany
- In my spare time I enjoy cooking and hiking.
- I was born in Yichang, Hubei, a key transportation hub on the Yangtze River. The city is known for its proximity to the Three Gorgets Dam, the world's largest hydroelectric power station.
- Many thanks to Nelson Liu for sharing the source code of this website!