Chao Zhang

I develop machine learning and data-centric solutions to build AI models for task support and decision making. My research is focused on two primary themes: (1) Data-Centric AI – How can we enhance AI models by addressing challenges in data quality and data scarcity? (2) Trustworthy AI – How can we make AI models more reliable and better aligned with human values? The ultimate goal is to rapidly develop more data-efficient and trustworthy AI systems that are customized for diverse domains in science and engineering. Currently, my group is actively working on the following topics:

• Large Language Models – Language model agents; Aligning LLMs to make them more trustworthy.
• Learning from Noisy Data – Fine-tuning language models with easy-to-obtain noisy labels.
• Uncertainty Quantification and Decision Making – Quantifying uncertainty in deep learning; Decision making under uncertainty.
• Spatiotemporal Dynamics and Design – Deep learning to simulate and forecast spatiotemporal dynamics (e.g., molecular simulation) for optimizing spatiotemporal systems.

On the application side, I am passionate about interdisciplinary research and enjoy developing data-driven solutions to accelerate scientific discovery through close collaboration with domain experts. The techniques I develop are motivated by applications in material science, biomedical science, transportation, and public health.

Acknowledgment: My work has been generously supported by research funding/gift from NSF (IIS CAREER-2144338, IIS-2106961, IIS-2008334), ONR MURI , Kolon, HomeDepot, ADP, and Adobe. My work has also been recognized by an NSF CAREER Award, a Facebook Faculty Award, an Amazon AWS Machine Learning Research Award, a Google Faculty Research Award, a Kolon Faculty Fellowship, an ACM SIGKDD Dissertation Runner-up Award, and several paper awards from IMWUT (UbiComp), ECML/PKDD, and ML4H.

Below are the main research projects at my group and some recent representative works:

• Large Language Models: LLMs are being extensively explored as versatile tools. Our objective is to make LLMs more capable and trustworthy for deployment in various domains. To achieve this, our research focuses on two main areas: 1) We are investigating LLM agents that can effectively utilize external tools (e.g., retrieval) and enhance their performance through interactions. 2) We are interested in identifying vulnerabilities within LLMs and scalably aligning them with natural language principles.
  • AdaPlanner: Adaptive Planning from Feedback with Language Models, preprint
  • ToolQA: A Dataset for LLM Question Answering with External Tools, preprint
  • Large Language Model as Attributed Training Data Generator: A Tale of Diversity and Bias, preprint
  • COCO-DR: Combating the Distribution Shift in Zero-Shot Dense Retrieval with Contrastive and Distributional Robust Learning, EMNLP 2022

• Learning from Noisy Data: Extracting information (e.g., entities, relations, events) from unstructured documents is a crucial task, but it often faces the challenge of lacking annotated data for model training. To address this issue, we combine pre-trained language models (e.g., BERT) with weakly-labeled data induced from labeling rules. We have developed techniques for fine-tuning language models with weak supervision and discovering rules from language models for interactive weak supervision.
  • DyGen: Fine-Tuning Language Models with Noisy Labels by Dynamics-Enhanced Generative Modeling, KDD 2023
  • Fine-Tuning Pre-trained Language Model with Weak Supervision, NAACL 2021
  • PRBoost: Prompt-Based Rule Discovery and Boosting for Interactive Weakly-Supervised Learning, ACL 2022
  • AcTune: Uncertainty-Aware Active Self-Training for Active Fine-Tuning of Language Models, NAACL 2022
  • ReSel: N-ary Relation Extraction from Scientific Text and Tables by Learning to Retrieve and Select, EMNLP 2022
  • Sparse Conditional Hidden Markov Model for Weakly Supervised Named Entity Recognition, KDD 2022
  • BERTifying Hidden Markov Models for Multi-Source Weakly Supervised Named Entity Recognition, ACL 2021
  • Text Classification Using Label Names Only: A Language Model Self-Training Approach, EMNLP 2020
  • Denoising Multi-Source Weak Supervision for Neural Text Classification, EMNLP 2020
  • Weakly-Supervised Neural Text Classification, CIKM 2018

• Uncertainty Quantification & Decision Making: Uncertainty-aware ML models are crucial for building trustworthy AI systems. Unfortunately, many deep learning models produce uncertainty-agnostic point estimates or miscalibrated distributions. To address this, we have been developing techniques (1) for quantifying uncertainty in deep learning models and (2) for exploiting uncertainty to inform downstream decision-making.
  • MUBen: Benchmarking the Uncertainty of Pre-Trained Models for Molecular Property Prediction, preprint
  • DF2: Distribution-Free Decision-Focused Learning, preprint
  • SMURF-THP: Score Matching-based UnceRtainty quantiFication for Transformer Hawkes Process, ICML 2023
  • End-to-end Stochastic Programming with Energy-based Model, NeurIPS 2022
  • UnfoldML: Cost-Aware and Uncertainty-Based Dynamic 2D Prediction for Multi-Stage Classification, NeurIPS 2022
  • Calibrated Language Model Fine-Tuning for In- and Out-of-Distribution Data, EMNLP 2020
  • CAMUL: Calibrated and Accurate Multi-view Time-Series Forecasting, WWW 2022
  • When in Doubt: Neural Non-Parametric Uncertainty Quantification for Epidemic Forecasting, NeurIPS 2021
  • SDE-Net: Equipping Deep Neural Networks with Uncertainty Estimates, ICML 2020
  • RDeepSense: Reliable Deep Mobile Computing Models with Uncertainty Estimations, UbiComp 2018

• Spatiotemporal Dynamics and Design: We study spatiotemporal dynamics and design in scientific and engineering applications. Our focus spans from macro-level systems such as transportation and epidemiology to micro-level systems such as molecules. Our mission is to develop cutting-edge machine learning methods to model and forecast the dynamics of these complex systems. This includes learning to simulate dynamics and spatiotemporal forecasting. Additionally, we investigate methods for intervening in these systems to achieve desired outcomes, using techniques such as end-to-end learning for optimization tasks and deep generative models for inverse design.
  • May the Force be with You: Unified Force-Centric Pre-Training for 3D Molecular Conformations, preprint
  • Autoregressive Diffusion Model for Graph Generation, ICML 2023
  • When Rigidity Hurts: Soft Consistency Regularization for Probabilistic Hierarchical Time Series Forecasting, KDD 2023
  • When in Doubt: Neural Non-Parametric Uncertainty Quantification for Epidemic Forecasting, NeurIPS 2021
  • T-GCN: A Temporal Graph Convolutional Network for Traffic Prediction, IEEE ITS 2020
  • A Linear Time Approach to Computing Time Series Similarity based on Deep Metric Learning, TKDE 2020
  • DeepMove: Predicting Human Mobility with Attentional Recurrent Networks, WWW 2018

• 2022 ML4H Outstanding Paper Award
• 2022 NSF Career Award
• 2021 Facebook Faculty Research Award
• 2021 Kolon Faculty Fellowship
• 2020 Amazon AWS Machine Learning Research Award
• 2020 Google Faculty Research Award
• 2019 ACM SIGKDD Dissertation Award Runner-up
• 2018 ACM IMWUT Distinguished Paper Award
• 2015 ECML/PKDD Best Student Paper Runner-up Award
• 2013 Chiang Chen Overseas Graduate Fellowship

(* denotes equal contribution)

• SDE-Net: Efficient uncertainty estimation for deep neural networks
• CHMM: BERT-conditional hidden Markov model for multi-source weakly-supervised learning
• COSINE: Language model fine-tuning with weak supervision
• BOND: Distantly-supervised named entity recognition
• STEAM: Automatic taxonomy expansion
• TaxoGen: Unsupervised topic taxonomy construction from text corpus
• WestClass: Weakly-supervised text classification
• GeoBurst: Unsupervised spatiotemporal event detection

• 2023 Fall: CSE8803-DLT: Deep Learning for Text Data
• 2023 Spring: CX4240: Introduction to Computational Data Analysis
• 2022 Spring: CX4240: Introduction to Computational Data Analysis
• 2021 Fall: CSE8803-DLT: Deep Learning for Text Data
• 2021 Spring: CX4240: Introduction to Computational Data Analysis
• 2020 Spring: CX4240: Introduction to Computational Data Analysis
• 2020 Fall: CSE8803-DLT: Deep Learning for Text Data
• 2019 Fall: CSE8803-DLT: Deep Learning for Text Data
• 2019 Spring: CX4240: Introduction to Computational Data Analysis

Prospective students: I am always looking for strong and motivated students to join our group. If you are interested in working with me, you can either email me or fill out this form.

Current:

• Rui Feng: Ph.D. Student in CS
• Lingkai Kong: Ph.D. Student in CSE
• Yinghao Li: Ph.D. Student in ML
• Agam A. Shah: Ph.D. Student in ML (co-advised with Sudheer Chava)
• Phillip Si: Ph.D. Student in CSE (co-advised with Peng Chen)
• Haotian Sun: Ph.D. Student in ML (co-advised with Bo Dai)
• Haorui Wang: Ph.D. Student in CSE
• Kuan Wang: Ph.D. Student in CSE
• Yue Yu: Ph.D. Student in CSE
• Rongzhi Zhang: Ph.D. Student in ML
• Yuchen Zhuang: Ph.D. Student in ML
• Jacob Wessell: M.S. student in CS
• Wenhao Mu: M.S. student in CS
• Shangqing Xu: M.S. student in CS

Alumni:

• Yanbo Xu: Ph.D., 2023 (First Employment: Microsoft Research)
• Binghong Chen: Ph.D., 2023 (co-advised with Prof. Le Song, First Employment: Citadel Capital)
• Pranav Shetty: Ph.D., 2023 (JP Morgan AI Ph.D. Fellowship, co-advised with Prof. Rampi Ramprasad, First Employment: JP Morgan Chase)
• Vidit Jain: M.S. Student in CS
• Mukund Rungta: M.S. Student in CS
• Junyang Zhang: M.S. Student in CS
• Piyush Patil: M.S. Student in CS
• Mengyang Liu: M.S. Student in CSE
• Isaac Rehg: M.S. in CS
• Wendi Ren: M.S. in CSE
• Ruijia Wang: M.S. in CSE
• Yi Rong: Visiting Ph.D. Student