Wenting Zhao

Ph.D. Candidate
Cornell University
[wzhao@cs.cornell.edu] [cv] [google scholar] [twitter]

I work on natural language processing and machine learning, with a primary focus on reasoning. My work explores reasoning in two directions:

  • Modeling: I develop techniques that can accurately reason in real-world scenarios. Here, I combine neural models with probabilistic methods to train strong reasoning models without reliance on labeled data.
  • Evaluation: I create benchmarks that reliably reflect the performance of LMs when deployed in the real world. To achieve this goal, I devise methods to gather large-scale data from user interactions observed in natural environments and focus evaluation on real-world use cases.

I am advised by Claire Cardie and Sasha Rush.

I’m on the faculty job market for 2024-2025 (and looking a postdoc position for a gap year)! Here are my research statement, teaching statement, and DEI statement. Please reach out if you have an opening in your department.

News

  • Oct 2024: I will be attending COLM 2024 in Philadelphia. Please reach out if you want to chat!
  • Sep 2024: We released Commit0, an AI coding challenge. Can your agent generate a software library from scratch?
  • Aug 2024: I will co-organize the second workshop on Natural Language Reasoning and Structured Explanations Workshop co-located with ACL 2024. Please consider submitting to our workshop!
  • Apr 2024: I will be attending ICLR 2024 in Vienna and NAACL 2024 in Mexico City. Please reach out if you want to chat!
  • Jan 2024: I was awarded the AI2 Outstanding Intern Award! Thanks to Alane, Lorraine, and Yejin :)
  • Dec 2023: I co-chaired the Bird-of-a-Feather Session on "Recipes in Building Language Reasoners" with Yiyuan Li!
  • Nov 2023: We released WildChat -- 650K conversations between users and ChatGPT!
  • Oct-Nov 2023: I presented WildChat at UCI, UCSB NLP, USC/ISI NLG Seminar, and NLP with Friends! If you missed my talk, here are my slides and a talk recording.

Selected Publications

I Could've Asked That: Reformulating Unanswerable Questions
Wenting Zhao, Ge Gao, Claire Cardie, Alexander M. Rush.
EMNLP 2024

WildChat: 1M ChatGPT Interaction Logs in the Wild
Wenting Zhao, Xiang Ren, Jack Hessel, Claire Cardie, Yejin Choi, Yuntian Deng.
ICLR 2024 (spotlight)

UNcommonsense Reasoning: Abductive Reasoning about Uncommon Situations
Wenting Zhao, Justin T Chiu, Jena D. Hwang, Faeze Brahman, Jack Hessel, Sanjiban Choudhury, Yejin Choi, Xiang Lorraine Li, Alane Suhr.
NAACL 2024

HOP, UNION, GENERATE: Explainable Multi-hop Reasoning without Rationale Supervision
Wenting Zhao, Justin Chiu, Claire Cardie, Alexander M Rush.
EMNLP 2023

Abductive Commonsense Reasoning Exploiting Mutually Exclusive Explanations
Wenting Zhao, Justin Chiu, Claire Cardie, Alexander M Rush.
ACL 2023

Contact

Wenting Zhao
wzhao at cs.cornell.edu
Cornell Tech