About Me

I am Zhaoqi “ZQ” Cheng, a PhD candidate at the Questrom School of Boston University. I obtained my bachelor’s degree at Tsinghua University and master’s degree at Carnegie Mellon University. I work with my advisor Dokyun Lee at the BIT Lab. Here is my CV.

Research Interests

I study the development and the responsible use of generative AIs for managerial and societal insights.

My work primarily delves into the interpretable representation and exploration of important business constructs in text corpora, with particular emphasis on patent innovation. Within this research theme, I have built InnoVAE, a generative model that maps and augments firm innovation, along with a data-driven approach for identifying tech-business zones to scrutinize antitrust implications of mergers and acquisitions.

Beyond methodological pursuits, my research interests extend to harnessing generative AIs for social good. Current projects in this domain include an ethical examination of deceptions in conversational AI, and the characterization of misinformation interventions leveraging Large Language Model agents.

Research Projects

Towards Cognition-Aware Language Agents: An Analysis in the Context of Misinformation
Zhaoqi Cheng, Yusen Wu, Dokyun Lee
To be presented at WISE 2023

InnoVAE: Generative AI for Understanding Patents and Innovation
Zhaoqi Cheng, Dokyun Lee, Prasanna Tambe
Presented at CIST 2020, INFORMS WDS 2020 and WISE 2020
2nd Round at Management Science [SSRN]

M&A and Innovation: A New Classification of Patents
Zhaoqi Cheng, Ginger Zhe Jin, Mario Leccese, Dokyun Lee, Liad Wagman
Presented at ASSA 2023 Annual Meeting
Available at AEA Papers and Proceedings [AEA]

Guided Diverse Concept Miner: Interpretable Deep Learning for Text Exploration
Dokyun Lee, Zhaoqi Cheng, Emaad Manzoor, Chengfeng Mao
3rd Round at Information Systems Research [SSRN][Poster]

The less tangled web: Why AI should practice deceit less than humans
Tae Wan Kim, Joy Lu, Kyusong Lee, John Hooker, Zhaoqi Cheng, Yanhan Tang
Presented at SBE Annual Meeting 2021 [arXiv]