Research

At OUNLP, we explore the frontier of natural language processing and machine learning with the goal of building intelligent, trustworthy, and practical AI systems. Our work spans multi-party, multi-modal dialogue and discourse analysis (in domains like education and mental health), agentic models and domain-specific “world” models for human-AI teaming, efficient structured-prediction and symbolic methods to augment neural networks, and the robust deployment and evaluation of trustworthy AI.

Highlighted

A Mamba-type of deep state space model for reservoir release simulation with a large-scale verification over 441 dams across CONUS

Jiaorui Zhang, Haowen Yue, Milad Basirifard, Jie Cao, Tiantian Yang

Journal of Hydrology  ·  01 Dec 2025  ·  doi:10.1016/j.jhydrol.2025.134145

reservoir-simulation deep-learning water-management interpretability hydrology

All

2025

A Mamba-type of deep state space model for reservoir release simulation with a large-scale verification over 441 dams across CONUS

Jiaorui Zhang, Haowen Yue, Milad Basirifard, Jie Cao, Tiantian Yang

Journal of Hydrology  ·  01 Dec 2025  ·  doi:10.1016/j.jhydrol.2025.134145

reservoir-simulation deep-learning water-management interpretability hydrology

“Understanding Robustness Lottery”: A Geometric Visual Comparative Analysis of Neural Network Pruning Approaches

Zhimin Li, Shusen Liu, Xin Yu, Kailkhura Bhavya, Jie Cao, James Daniel Diffenderfer, Peer-Timo Bremer, Valerio Pascucci

IEEE Transactions on Visualization and Computer Graphics  ·  01 Sep 2025  ·  doi:10.1109/TVCG.2024.3514996

neural-network-pruning robustness visual-analysis geometric-comparative-analysis

Towards Actionable Pedagogical Feedback: A Multi-Perspective Analysis of Mathematics Teaching and Tutoring Dialogue

Jannatun Naim, Jie Cao, Fareen Tasneem, Jennifer Jacobs, Brent Milne, James Martin, Tamara Sumner

International Educational Data Mining Society  ·  12 Jul 2025  ·  doi:10.5281/zenodo.15870176

pedagogical-feedback mathematics-education talk-moves discourse-analysis tutoring-dialogue

MatterChat: A Multi-Modal LLM for Material Science

Yingheng Tang, Wenbin Xu, Jie Cao, Weilu Gao, Steve Farrell, Benjamin Erichson, Michael W. Mahoney, Andy Nonaka, Zhi Yao

arXiv  ·  01 Jan 2025  ·  doi:10.48550/arXiv.2502.13107

materials-science multi‑modal-llm structure‑aware-modeling property-prediction interatomic-potential

AQUAH: Automatic Quantification and Unified Agent in Hydrology

Songkun Yan, Zhi Li, Siyu Zhu, Yixin Wen, Mofan Zhang, Mengye Chen, Jie Cao, Yang Hong

arXiv  ·  01 Jan 2025  ·  doi:10.48550/arXiv.2508.02936

hydrologic-modeling language-agents vision-llm automatic-data-retrieval simulation

Rethinking On-policy Optimization for Query Augmentation

Zhichao Xu, Shengyao Zhuang, Xueguang Ma, Bingsen Chen, Yijun Tian, Fengran Mo, Jie Cao, Vivek Srikumar

arXiv  ·  01 Jan 2025  ·  doi:10.48550/arXiv.2510.17139

query-augmentation information-retrieval large-language-models reinforcement-learning

2024

Enhancing Talk Moves Analysis in Mathematics Tutoring through Classroom Teaching Discourse

Jie Cao, Abhijit Suresh, Jennifer Jacobs, Charis Clevenger, Amanda Howard, Chelsea Brown, Brent Milne, Tom Fischaber, Tamara Sumner, James H. Martin

arXiv  ·  01 Jan 2024  ·  doi:10.48550/arxiv.2412.13395

mathematics-tutoring saga22-dataset pretraining dialogue-context