Yuanpei Cao

About Me

I am a staff Machine Learning Engineer at Intelligent Support Platform team (San Francisco), where I work on NLP models for Airbnb Chatbot and computer vision models for search ranking, fraud detection, and marketing, etc.

In August 2016, I received a Ph.D. from the Department of Mathematics at University of Pennsylvania, specializing in Applied Mathematics and Computational Science . My advisor was Prof. Hongzhe Li, and my thesis was titled Statistical Methods for High Dimensional Count and Compositional Data with Applications to Microbiome Studies.

Resume: https://github.com/yuanpeicao/yuanpeicao.github.io/blob/master/resume.pdf

Linkedin: https://www.linkedin.com/in/yuanpei-cao-792b103b/

Google Scholar: https://scholar.google.com/citations?user=NeDKHAcAAAAJ&hl=en

Kaggle homepage: https://www.kaggle.com/caoyuanpei

Github: https://github.com/yuanpeicao

Education

• Ph.D. in Applied Mathematics and Computational Science, University of Pennsylvania (2011-2016)
• M.A. in Statistics, Wharton School, University of Pennsylvania (2011-2016)
• B.S. in Applied Mathematics, Fudan University (2007-2011)

Patent

• Tree Structure-Based Smart Inter-Computing Routing Model, Ref 26887-41473, 06/18/2020 [link]

• Voice Synchronized Visual Interface, PCT Application series number PCT/CN2023/091952, 05/04/2023 [link]

• Large Language Model-based Customized Conversational Message, PCT Application series number PCT/CN2023/127630, 10/31/2023 [link]

• Large Language Model-based Interactive Story-based Authentication in IVR, PCT Application series number PCT/CN2024/124145, 10/11/2024 [link]

Public Tech Blog

• WIDeText: A Multimodal Deep Learning Framework (with Zhang, W. and Zhao, M) [link]
• When a Picture Is Worth More Than Words (with Ulammandakh, B. Wang, H. and Tony, H) [link]

Kaggle Competition

• New York City Taxi Fare Prediction , Hosted by Google and R Studio, Sept. 2018
  Final Leaderboard Rank: 9/1488 (top 1%)
• Google Analytics Customer Revenue Prediction , Hosted by Google Cloud and Coursera, Feb. 2019
  Final Leaderboard Rank: 11/1089 (top 1%, Gold Medal)

Papers

• Enhancing the accessibility of regionalization techniques through large language models: a case study in conversational agent guidance (with Feng, X)
  International Journal of Geographical Information Science (2024), 1–24. https://doi.org/10.1080/13658816.2024.2415439. [PDF]
• Multi-sample Estimation of Bacterial Composition Matrix in Metagenomics Data (with Zhang, A. and Li, H)
  Biometrika (2020), 107, 75-92. [PDF]
• MOTO: A Multi-Objective Trajectory Optimization Method for Finding Sequential Activity Locations for Multiple Moving Objects along Road Networks (with Feng, X., Wang, S., Gao, S., and Murray, A.T)
  Environment and Planning B: Urban Analytics and City Science (2020). https://doi.org/10.1177/2399808320913300. [PDF]
• Large Covariance Estimation for Compositional Data via Composition-Adjusted Thresholding (with Lin, W. and Li, H)
  Journal of the American Statistical Association (JASA) (2019) [PDF]
• Two-sample Mean Tests for High Dimensional Compositional Data (with Lin, W. and Li, H)
  Biometrika (2018), 105, 115–132. [PDF]

Working Papers

• Sparse High-Dimensional Precision Matrix Estimation for Compositional Data. (with Ma, R. and Li, H)
  Submitted.
• Multi-sample estimation of centered log-ratio matrix in microbiome studies (with Li, Y. and Li, H)
  Submitted. [link]