Kaiwen Wu

About

I am currently a research scientist at Meta. I did my PhD in Computer and Information Science at the University of Pennsylvania, advised by Jacob Gardner, where I worked on probabilistic machine learning. Before that, I completed my MMath in Computer Science at the University of Waterloo.

Research

* indicates equal contribution.

• Self-Driving Datasets: From 20 Million Papers to Nuanced Biomedical Knowledge at Scale
  Haydn Jones, Yimeng Zeng, Alden Rose, Li S. Yifei, Yining Huang, Kaiwen Wu, Jiaming Liang, Maggie Ziyu Huan, Yoseph Barash, Cesar de la Fuente-Nunez, Osbert Bastani, Zachary Ives, Mark Yatskar, Jacob R. Gardner
  arXiv preprint arXiv:2605.07022, 2026
• Knowledge Gradient for Preference Learning
  Kaiwen Wu, Jacob R. Gardner
  arXiv preprint arXiv:2601.22335, 2026
• Mixed Likelihood Variational Gaussian Processes
  Kaiwen Wu, Craig Sanders, Benjamin Letham, Phillip Guan
  arXiv preprint arXiv:2503.04138, 2025
• Computation-Aware Gaussian Processes: Model Selection and Linear-Time Inference
  Jonathan Wenger, Kaiwen Wu, Philipp Hennig, Jacob R. Gardner, Geoff Pleiss, John P. Cunningham
  Advances in Neural Information Processing Systems (NeurIPS), 2024
• Understanding Stochastic Natural Gradient Variational Inference
  Kaiwen Wu, Jacob R. Gardner
  International Conference on Machine Learning (ICML), 2024
• Large-Scale Gaussian Processes via Alternating Projection [code]
  Kaiwen Wu, Jonathan Wenger, Haydn Jones, Geoff Pleiss, Jacob R. Gardner
  International Conference on Artificial Intelligence and Statistics (AISTATS), 2024
• A Fast, Robust Elliptical Slice Sampling Implementation for Linearly Truncated Multivariate Normal Distributions [code]
  Kaiwen Wu, Jacob R. Gardner
  Workshop on Bayesian Decision-Making and Uncertainty at NeurIPS 2024
• The Behavior and Convergence of Local Bayesian Optimization [code]
  Kaiwen Wu, Kyurae Kim, Roman Garnett, Jacob R. Gardner
  Advances in Neural Information Processing Systems (NeurIPS), 2023Spotlight
• On the Convergence of Black-Box Variational Inference [code]
  Kyurae Kim, Jisu Oh, Kaiwen Wu, Yian Ma, Jacob R. Gardner
  Advances in Neural Information Processing Systems (NeurIPS), 2023
• Variational Gaussian Processes with Decoupled Conditionals [code]
  Xinran Zhu, Kaiwen Wu, Natalie Maus, Jacob Gardner, David Bindel
  Advances in Neural Information Processing Systems (NeurIPS), 2023
• Practical and Matching Gradient Variance Bounds for Black-Box Variational Bayesian Inference
  Kyurae Kim, Kaiwen Wu, Jisu Oh, Jacob R. Gardner
  International Conference on Machine Learning (ICML), 2023Oral
• Local Bayesian Optimization via Maximizing Probability of Descent [code]
  Quan Nguyen*, Kaiwen Wu*, Jacob R. Gardner, Roman Garnett
  Advances in Neural Information Processing Systems (NeurIPS), 2022Oral
• Discovering Many Diverse Solutions with Bayesian Optimization [code]
  Natalie Maus, Kaiwen Wu, David Eriksson, Jacob Gardner
  International Conference on Artificial Intelligence and Statistics (AISTATS), 2022Notable Paper
• Newton-type Methods for Minimax Optimization [code]
  Guojun Zhang, Kaiwen Wu, Pascal Poupart, Yaoliang Yu
  Workshop on Beyond First-Order Methods in ML Systems at ICML 2021
• Stronger and Faster Wasserstein Adversarial Attacks [code]
  Kaiwen Wu, Allen Houze Wang, Yaoliang Yu
  International Conference on Machine Learning (ICML), 2020
• On Minimax Optimality of GANs for Robust Mean Estimation [code]
  Kaiwen Wu, Gavin Weiguang Ding, Ruitong Huang, Yaoliang Yu
  International Conference on Artificial Intelligence and Statistics (AISTATS), 2020

Notes

For self-reference only. Some take forever to finish.

• A Gentle Introduction to Variational Gaussian Processes
• Concepts in Smooth Manifolds and Differential Geometry
• Normalization in Gaussian Processes
• L1 Norm Projection

Miscellaneous

Reviewer for AAAI 2021, AISTATS 2021, ICML 2023, NeurIPS 2023, ICLR 2024, AISTATS 2024, ICML 2024, NeurIPS 2024, ICLR 2025, ICML 2025, NeurIPS 2025, AAAI 2026, ICLR 2026, ICML 2026.

Reviewer for TMLR and JMLR.

My Erdős number is 3.