I'm an AI researcher working on foundation models. Before starting my career as an AI researcher in industry, I earned my Ph.D. degree in Electrical and Computer Engineering (Machine Learning track) from the University of Michigan, Ann Arbor. My research focuses on LLMs, reinforcement learning, alignment, trustworthy ML, and AI for science. I'm open to collaboration with aspiring researchers and I frequently mentor students on AI research projects. If you find shared research interests and would like to discuss collaboration opportunities, please feel free to contact me via email!
The full list of my publications can be found on Google Scholar.
Graph Neural Prompting with Large Language Models [large language models, knowledge graphs]
Yijun Tian,
Huan Song,
Zichen Wang,
Haozhu Wang,
Ziqing Hu,
Fang Wang,
Nitesh V.Chawla,
Panpan Xu
AAAI, 2024
arxiv
A knowledge graph prompting method for large language models to improve their commonsense and biomedical reasoning performance.
A Review of Reinforcement Learning for Natural Language Processing, and Applications in Healthcare [reinforcement learning, LLMs]
Ying Liu,
Haozhu Wang,
Huixue Zhou,
Mingchen Li,
Yu Hou,
Sicheng Zhou,
Fang Wang,
Rama Hoetzlein,
Rui Zhang
under review, 2023
arxiv
A comprehensive review of reinforcement learning applied to NLP and its healthcare applications.
T3GDT: Three-Tier Tokens to Guide Decision Transformer for Offline Meta Reinforcement Learning [reinforcement learning, foundation model]
Zhe Wang,
Haozhu Wang,
Yanjun Qi
NeurIPS Workshop on Robot Learning, 2023
paper
We developed an hierarchical prompting method for transformer-based reinforcement learning models to enable efficient few-shot policy adaptation.
Latent skill discovery for chain-of-thought reasoning [LLMs, foundation model]
Zifan Xu,
Haozhu Wang,
Dmitriy Bespalov,
Yanjun Qi
NeurIPS Workshop on Robustness of Zero/Few-shot Learning in Foundation Models, 2023
paper
We developed an unsupervised method for discovering latent skills to guide the demonstration selection for in-context learning with large langauge models.
Reinforcement Learning-Enabled Environmentally Friendly and Multi-functional Chrome-looking Plating [AI for science, reinforcement learning]
Taigao Ma,
Anwesha Saha,
Haozhu Wang,
L. Jay Guo,
NeurIPS AI for Science Workshop, 2023, [Oral, selection rate: 10/150=6.7%] OpenReview
Using reinforcement learning, we designed and fabricated two multilayer thin film structures that can mimic the visual appearance of decorative chrome plating, serving as a environmentally friendly and multi-functional replacement.
OptoGPT: A Foundation Model for Inverse Design in Optical Multilayer Thin Film Structures [AI for science, foundation model]
Taigao Ma,
Haozhu Wang,
L. Jay Guo
under review, 2023
arXiv
We developed OptoGPT, the first foundation model for optical thin film structure inverse design. After being trained on a large dataset of 10 million optical thin film designs, OptoGPT demonstrates remarkable capabilities including: 1) autonomous global design exploration, 2) efficient designs for various tasks, 3) the ability to output diverse designs, and 4) seamless integration of user-defined constraints. We believe OptoGPT is a major leap towards accelerating optical science with foundation models.
Structural color generation: from layered thin films to optical metasurfaces [AI for science]
Danyan Wang,
Zeyang Liu,
Haozhu Wang,
Moxin Li,
L. Jay Guo,
Cheng Zhang
Nanophotonics, 2023
paper
Comprehensive survey of the structural color research field. I provided a discussion of applying machine learning to structural color device designs.
Dynamic prediction of work status for workers with occupational injuries: assessing the value of longitudinal observations [ML for healthcare]
Erkin Ötleş,
Jon Seymour,
Haozhu Wang,
Brian T Denton
Journal of the American Medical Informatics Association, 2022
paper
We developed a forecasting model to predict return-to-work after occupational injuries based on longitudinal claim data. The model may allow case managers to better allocate medical resources and help speed up patients' recover process.
NEUTRON: Neural Particle Swarm Optimization for Material-Aware Inverse Design of Structural Color [AI for science] Haozhu Wang,
L. Jay Guo
iScience, 2022
paper/
code
We propose a hybrid machine learning and optimization method that combines mixture density networks and particle swarm optimization for accurate and efficient structural color inverse design.
Benchmarking Deep Learning-based Models on Nanophotonic Inverse Design Problems [AI for science]
Taigao Ma,
Mustafa Tobah,
Haozhu Wang* ,
L. Jay Guo*
Opto-Electronic Science, 2022 (*: correspondence)
paper
We provide extensive benchmarking results on accuracy, diversity, robustness for commonly used deep learning models in nanophotonic inverse designs. The findings can help researchers select models that best suit their design problems.
An Analytical Method for Evaluating the Robustness of Photonic Integrated Circuits [AI for science]
Hanfa Song,
Haozhu Wang ,
Vien Van
Journal of Lightwave Technology, 2022
paper
We provide an analytical approach for evaluating the robustness of photonic integrated circuits. The method is verified by genetic algorithms.
Automated Optical Multi-layer Design via Deep Reinforcement Learning [AI for science, reinforcement learning] Haozhu Wang ,
Zeyu Zheng,
Chengang Ji,
L. Jay Guo
Machine Learning: Science and Technology, 2021
paper/
code/
abridged NeurIPS workshop version/
Training a novel sequence generation network with Proximal Policy Optimization for automatically discovering near-optimal optical designs.
COVID-19 Risk Scoring in Los Angeles County [ML for healthcare]
Litian Zhou, Wenxue Li, Zhangxing Bian, Yuxuan Cao, Xinyu Li, Weixiao Wang, Zixian Ma
Junhwan Kim, Zijin Chu, Yuxi Xie, Yueze Song, Chaoyi Wang, Ruopeng Wang, Linh Tran
Haozhu Wang*, L. Jay Guo*
RMDS COVID-19 Challenge, 2020 (*: correspondence)
report
Apply LSTM and LR models for spatio-temporal COVID-19 risk prediction.
Return to Work After Injury: A Sequential Prediction & Decision Problem [ML for healthcare]
Erkin Ötleş*
Haozhu Wang *,
Suyanpeng Zhang,
Brian Denton,
Jenna Wiens,
Jon Seymour
Machinet Learning for Healthcare (clinical abstract), 2019 (*: equal contribution)
report
Apply Q-learning to insurance claim data to learn near-optimal dynamic treatment regimes.
Expert-yielded-estimates regularizer for incorporating expert knowledge into linear models.
Learning to Share: Simultaneous Parameter Tying and Sparsification in Deep Learning [model compression]
Dejiao Zhang*,
Haozhu Wang *,
Mario A.T. Figueiredo,
Laura Balzano
ICLR, 2018 (*: equal contribution)
paper/
code
Group-ordered-weighted lasso (GrOWL) for deep model compression.
Teaching
Mentored > 30 graduate and undergraduate students on AI research projects.