Resumé
Education
Physical Sciences Division, The University of Chicago (Sep. 2021 - Oct. 2025 Expected)
Ph.D. student in Computer Science, GPA: 3.98
Focus: Machine Learning for Science, Bayesian Deep Learning, Bayesian Optimization (BO)
Physical Sciences Division, The University of Chicago (Sep. 2019 - Dec. 2020)
M.S. in Computer Science, GPA: 3.98
Focus: Full Stack Development, Machine Learning, High-Performance Computation
School of Data and Computer Science, Sun Yat-Sen University (Aug. 2015 - June. 2019)
B.S. in Information and Computing Science, GPA: 4.0/4.0, Ranking: 2/58
Focus: Machine Learning, High-Performance Computer
Honors & Services
- Reviewer for Top Machine Learning Conferences:
- NeurIPS 2023 (Top reviewer)
- ICML 2022-2024
- ICLR 2023-2024
- IJCAI 2024
- Scholarship:
- China National Scholarship 2017-2018 (1/55)
- Sun Yat-Sen Fellowship 2016-2019 (5/55)
- Crerar Fellowship 2021-2022
Research Interests
My research interests center on learning and decision-making under conditions of uncertainty. Specifically, I am intrigued by the application of novel deep learning and statistical learning methods to diverse fields, including protein design, materials science, content distribution, and in-context learning for large language models. By performing uncertainty quantification on modern models, I aim to make sample-efficient decisions that proactively gather new data and propose incoming candidates or content, thereby optimizing application-specific objectives.
Publications
-
(𝛼 − 𝛽) Han, Minbiao, Fengxue Zhang, and Yuxin Chen. “No-Regret Learning of Nash Equilibrium for Black-Box Games via Gaussian Processes.” In The 40th Conference on Uncertainty in Artificial Intelligence, 2024.
-
Zhang, Fengxue, Jialin Song, James C. Bowden, Alexander Ladd, Yisong Yue, Thomas Desautels, and Yuxin Chen. “Learning regions of interest for Bayesian optimization with adaptive level-set estimation.” In International Conference on Machine Learning, pp. 41579-41595. PMLR, 2023.
-
Zhang, Fengxue, Thomas Desautels, and Yuxin Chen. “Robust Multi-fidelity Bayesian Optimization with Deep Kernel and Partition.” In the 28th International Conference on Artificial Intelligence and Statistics.
-
Li, Diantong, Fengxue Zhang, Chong Liu, and Yuxin Chen. “ Constrained Multi-objective Bayesian Optimization through Optimistic Constraints Estimation.” In the 28th International Conference on Artificial Intelligence and Statistics.
-
Zhang, Fengxue, Zejie Zhu, and Yuxin Chen. “Finding Interior Optimum of Black-box Constrained Objective with Bayesian Optimization.” In NeurIPS 2024 Workshop on Bayesian Decision-making and Uncertainty.
-
Zhang, Fengxue, Yair Altas, Louise Fan, Kaustubh Vinchure, Brian Nord, and Yuxin Chen. “Design of Physical Experiments via Collision-free Latent Space Optimization.” In NeurIPS 2020 Workshop on Machine Learning and the Physical Sciences.
Selected Projects
In-context-learning of LLM through Bayesian Optimization
University of Chicago, Sep. 2023 - Present
- Developed the fixed set selection category by modeling the LLM and its in-context examples as a black-box model
- Applied Bayesian Optimization (BO) to select the best set of permutations for different task types
- Applied random embedding, sparse kernel, and trust-region identification techniques to overcome the curse of high dimensionality
Phosphate Sensor Calibration with Deep Learning
University of Chicago & Argonne National Lab, Sep. 2023 - Oct. 2024
- Developed a novel preprocessing pipeline to calibrate cross-device variation together with variable selection and robust scaling
- Developed a two-branch neural network composed of both 1D and 2D convolutional neural to fuse multiple information sources
- Identified the optimized response window and predicted the true concentration of Phosphate in fertilizer solutions
Big Data Equity Market Analysis System
University of Chicago, Sep. 2020 - Dec. 2020
- Implemented the big data system using Lambda architecture, with a self-implemented ZeroMQ-style message queue and front-end
- Implemented distributed storage of the information in Google File System and optimized the Sharpe Ratio of the portfolio using Spark
- Fine-tuned and distilled the Large Language Models using Information Bottleneck to process information
Distributed Crawling System
Sun Yat-Sen University, May. 2017 - Sep. 2017
- Developed the crawling system based on the Scrapy-Redis framework
- Employed Bloom Filter to improve storage and communication
- Won the 3rd prize in the China Software Cup Software Design Competition 2017
Skills
- Domain: (Large-scale) Machine Learning, Big data (Lambda architecture), Distributed System
- Languages: Python, C/C++, MATLAB, Shell Script, Java, SQL, Go
- Frameworks/Tools: Pytorch, BoTorch, TensorFlow, MATLAB Neural Network Toolbox, Git