Steeve Huang


Co-Founder, Rosetta.ai


Kung-Hsiang (Steeve) Huang is a first-year PhD student at the University of Illinois, Urbana-Champaign. He is part of the BLENDER Lab, led by Professor Heng Ji, where his works mainly focus on fact checking and fake news detection. Prior to joining UIUC, he obtained his master’s degree at the University of Southern California. Under the supervision of Professor Nanyun Peng, he obtained rigorous training in conducting NLP research. His research mainly focuses on information extraction. Particularly, he has worked on knowledge-enhanced event extraction, document-level event extraction, and zero-shot cross-lingual transfer for event extraction. In addition, he is a co-founder of Rosetta.ai, in charge of the core Deep Learning-based recommender system that empowers our clients with over millions of online shoppers.


  • Fact Checking
  • Fake News Detection
  • Knowledge Reasoning
  • Information Extraction


  • PhD in Computer Science, 2021 -

    University of Illinois at Urbana-Champaign

  • MSc in Computer Science, 2020 - 2021

    University of Southern California

  • BEng in Computer Science, 2014 - 2018

    Hong Kong University of Science and Technlogy

  • Exchange Program, 2016

    Georgia Institute of Technology

Recent News

  • [Nov 2022] Will give a tutorial at AACL 2022!
  • [Aug 2022] Will give a tutorial at KDD 2022!
  • [Apr 2022] Serve as a reviewer for JAIR.
  • [Oct 2021] Serve as a reviewer for ACL Rolling Review.
  • [Aug 2021] One paper accepted to EMNLP 2021!
  • [Aug 2021] Moved from LA to Champaign ✈️.



Research Assistant

University of Illinois at Urbana-Champaign

Aug 2021 – Present Illinois, United States

Research Assistant

Information Sciences Institute

Jan 2020 – Jul 2021 California, United States

Chief Technology Officer


Aug 2018 – Aug 2021 Taipei, Taiwan

Machine Learning Engineer

Industrial Technology Research Institute

May 2018 – Aug 2018 Hsinchu, Taiwan

Junior Data Scientist


Feb 2018 – May 2016 Hong Kong

Recent Publications

Faking Fake News for Real Fake News Detection: Propaganda-loaded Training Data Generation
Document-level Entity-based Extraction as Template Generation
EventPlus: A Temporal Event Understanding Pipeline
Efficient End-to-end Learning of Cross-event Dependencies for Document-level Event Extraction
Biomedical Event Extraction with Hierarchical Knowledge Graphs


Dean’s Scholarship

Merit-based scholarship issued on admission.

4th Place, 2019 ACM RecSys Challenge

Build session-based, context-aware recommender systems for hotel metasearch.

Silver Medal, The 2nd YouTube-8M Video Understanding Challenge

Build a compact (< 1GB) video classification model for video understanding.

Competitions Master

Silver Medal, Avito Demand Prediction Challenge

Predict the demand for an online advertisement based on natural language metadata and context.

Bronze Medal, TalkingData AdTracking Fraud Detection Challenge

Detect fraudulent click traffic for mobile app ads.

Final List, IT Challenge

Create an innovative Chatbot solution.

Bronze Medal, 2018 Data Science Bowl

Create a model to identify the nuclei from an image, evaluated at the pixel-level.

Gold Medal, Toxic Comment Classification Challenge

Build multi-headed models that’s capable of detecting different types of of toxicity like threats, obscenity, insults, and identity-based hate.

Bronze Medal, Recruit Restaurant Visitor Forecasting

Construct a regression model for predicting the number of visitors a restaurant will receive.

Silver Medal, Statoil/C-CORE Iceberg Classifier Challenge

Build a classifier to distinuish between ship and iceberg from remotely shot images.

First Place, Smart City in Hong Kong

Second Place, Public Welfare Datathon (公益数据骇客马拉松)

First Runner-up, Imagine Cup Hong Kong Final

Reaching Out Award Scholarship

University Full Scholarship