Rochana Chaturvedi

prof_pic.jpg

I am a Postdoctoral Fellow at Argonne National Laboratory, where I develop and apply large language models for scientific applications. My work aims to advance scientific discovery and societal well-being through language technologies with applications across healthcare, social science, and infrastructure resilience.

I earned my Ph.D. in Computer Science from the University of Illinois Chicago, where I developed novel methods to extract and reason about temporal relationships in clinical text. My dissertation integrated graph neural networks, language models, and external knowledge bases to enable early detection of Type 2 Diabetes. I was fortunate to have Prof. Barbara Di Eugenio as my advisor and was grateful to receive mentorship from Prof. Elena Zheleva and Prof. Sourav Medya.

In social sciences, I employ computational approaches to explore societal dynamics and use language models to understand and model aspects of human behavior. I tackle challenges such as identity inference, social media polarization and its mitigation, and covert gender biases in large language models with their labor market impact.

Before my doctoral studies, I served as an Assistant Professor of Computer Science at Keshav Mahavidyalaya, University of Delhi, and worked as an Associate Software Engineer at Objective Systems Integrators. I hold Master’s degrees in Computer Science and Computer Applications, and a Bachelor’s degree in Physics.

News

Jul 10, 2025 🎓 Successfully defended my Ph.D. at UIC with the dissertation “Temporal Reasoning in Clinical Narratives: From Information Extraction to Early Disease Detection”. I am deeply grateful to my advisor, committee, collaborators, and my family.
May 15, 2025 🎉 Excited to share that our paper, “Temporal Relation Extraction in Clinical Texts: A Span-based Graph Transformer Approach,” has been accepted to ACL 2025! 📄
May 02, 2025 🚨 Our research on gender bias in open-source AI models was featured in The Register!
Jan 23, 2024 :confetti_ball: Thrilled to share that our paper Bridging or Breaking: Impact of Intergroup Interactions on Religious Polarization has been accepted at the ACM Web Conference 2024! Also presenting my thesis research “Temporal Knowledge Graph Extraction and Modeling across Multiple Documents for Health Risk Prediction” at the PhD symposium! :hourglass: Big thanks to NSF for the conference travel award! :raised_hands:
Nov 05, 2023 Our paper Sequential Representation of Sparse Heterogeneous Data for Diabetes Risk Prediction is accepted in IEEE International Conference on Bioinformatics and Biomedicine (BIBM) 2023. Thanks NSF for the conference travel grant! :tada:
Jul 12, 2023 :newspaper: Our findings from It’s All in the Name: A Character-Based Approach to Infer Religion are covered by popular news outlet: The Hindu :wave:
Sep 06, 2022 Presenting a lightning talk at NSF-NIH Smart and Connected Health workshop as a student PI. A great experience!