Xingchen (Cedric) Xu

Xingchen (Cedric) Xu

Researcher

Research

“*” denotes first-author or co-first-author papers. “+” denotes the first author’s inaugural first-authored paper in the PhD program (some thoughts about writing the first paper).

AI Economics (Human-AI interaction and AI-AI interaction)

  1. *Sooner but Weaker: The Dual Impact of Human Intervention on Algorithmic Collusion
    • Preliminary version appears in BizAI 2026 (UTD), CIST 2026 (San Francisco).
    • [Multi-agent Game], [Structural Estimation], [Lab Experiment], [Reinforcement Learning], [Analytical Model]
  2. *Depersonalizing Search on Social Media: A Large-Scale Field Experiment
    • Preliminary version appears in DSI 2026 (San Francisco).
    • Dissertation Chapter 3.
    • [Field Experiment], [Analytical Model], [Reduced-form Analyses]
  3. *Algorithmic Collusion or Competition
    • INFORMS ISS Cluster Best Paper Award, 2025.
    • Preliminary version appears in WDS 2023 (Phoenix), BizAI 2025 (UTD), POMS 2025 (Atlanta), INFORMS 2025 (Atlanta).
    • Dissertation Chapter 2.
    • [Multi-agent Game], [Structural Estimation], [Search Model], [Reinforcement Learning]
  4. *Generate the Future of Work through AI
    • WITS Best Paper Runner Up, 2023.
    • Preliminary version appears in Biz&GenAI 2023 (Wharton@San Francisco), WDS 2023 (Phoenix), AIMLBA 2023 (Temple@Philadelphia), WITS 2023 (Hyderabad, India), AI&Workplace 2024 (Wharton Online), CMIC 2024 (Zhengzhou, China), INFORMS 2024 (Seattle), CIST 2024 (Seattle), POMS 2025 (Atlanta).
    • Dissertation Chapter 1.
    • [Reduced-form Analyses]
  5. +Learning to Adopt Generative AI
    • Conditionally accepted by Information Systems Research.
    • DSI Best Interdisciplinary Paper Award, 2025.
    • Preliminary version appears in Biz&GenAI 2024 (Wharton@San Francisco), INFORMS 2024 (Seattle), CIST 2024 (Seattle), AIMLBA 2024 (Yale@New Haven), POMS 2025 (Atlanta).
    • [Structural Estimation], [Learning Model], [Reduced-form Analyses]
  6. +How People Use Generative AI in the Organization
    • Preliminary version appears in INFORMS 2026 (San Francisco), DSI 2026 (San Francisco).
    • [Structural Estimation], [Learning Model], [Reduced-form Analyses]
  7. Multimodal AI Feedback Improves Teaching Quality and Student Engagement Nationwide But Unevenly
    • Preliminary version appears in TRICS 2026 (Tulane), INFORMS 2026 (San Francisco).
    • [Reduced-form Analyses]
  8. *When Content is Goliath and Algorithm is David: The Style and Semantic Effects of Generative Search Engine
    • Preliminary version appears in Biz&GenAI 2023 (Wharton@San Francisco), INFORMS 2023 (Phoenix), Biz&AI 2024 (UTD@Dallas).
    • [Reduced-form Analyses]

Information Economics (Information and Platform Design/Algorithms)

  1. *Social Audience Size as a Reference Point: Evidence from a Field Experiment
    • Management Science, 2026.
    • Preliminary version appears in POMS 2022 (Virtual), CIST 2022 (Indianapolis), BBTC 2023 (Miami).
    • [Field Experiment], [Reduced-form Analyses]
  2. +Reputation Farming: A Structural Model of Freelancers’ Bidding Strategy
    • Preliminary version appears in INFORMS 2025 (Atlanta).
    • [Structural Estimation], [Dynamic Discrete Choice Model], [Reduced-form Analyses]
  3. *Learning Group Dynamics and Engagement Trap
    • Preliminary version appears in CIST 2022 (Indianapolis).
    • [Structural Estimation], [Hidden Markov Finite-Group Game], [Reinforcement Learning]
  4. +Same Space, Different Sparks
    • Preliminary version appears in INFORMS 2025 (Atlanta), CIST 2026 (San Francisco), INFORMS 2026 (San Francisco), DSI 2026 (San Francisco).
    • [Structural Estimation], [Point Process Model], [Reinforcement Learning]
  5. Political Consumerism and the Emergence of Rare Information on User-Generated Content Platforms
    • Forthcoming in Management Science.
    • [Reduced-form Analyses]
  6. Enhancing User Privacy Through Ephemeral Sharing Design: Experimental Evidence from Online Dating
    • Information Systems Research, 2025.
    • ICIS Best Paper in IS Design, Development and Project Management, 2021.
    • Preliminary version appears in CODE 2020 (MIT@Boston), WISE 2020 (Virtual), HCISS 2021 (Virtual), DSI 2021 (Atlanta), CSWIM 2021 (Virtual), ICIS 2021 (Austin).
    • [Field Experiment], [Lab Experiment], [Reduced-form Analyses]
  7. The Identity Tax: How Group Identity Disclosure Disproportionately Affects Online Discussions Across Hierarchies
    • Preliminary version appears in CIST 2023 (Phoenix).
    • [Reduced-form Analyses]

(Open-Source) LLM/VLM Building and Evaluation

  1. Open-Source Multimodal Moxin Models with Moxin-VLM and Moxin-VLA
  2. 7B Fully Open Source Moxin-LLM – From Pretraining to GRPO-based Reinforcement Learning Enhancement
  3. PhyGround: Benchmarking Physical Reasoning in Generative World Models