Sarah M. Preum | Department of Computer Science

Biography

I am a scientist working at the intersection of Natural Language Processing, Human–AI Interaction, and Computational Health. My research explores how AI can become a more effective communication partner in high-stakes settings, ultimately helping people exchange information, navigate uncertainty, and make informed decisions with less cognitive burden. I am particularly interested in socio-technical AI systems that are adaptive, equitable, and grounded in how people communicate and make decisions in the real world.

At the Pervasive Intelligent Systems Lab (PersistLab), my students and I develop human-centered NLP methods, systems, and evaluation frameworks for patient–provider messaging, online health communities, and other communication-intensive settings. Our work combines large-scale behavioral and clinical data with interdisciplinary collaboration and engagement with patients, clinicians, and communities in the USA and the Global South. We also prioritize reproducible and open research and have released more than a dozen public datasets and resources. To date, we have authored about 60 peer-reviewed publications in leading venues spanning natural language processing, human–computer interaction, artificial intelligence, and computational health, including ACL, EMNLP, AAAI, CHI, IMWUT, ICWSM, and ACM Computing Surveys.

My research has been recognized with a Google Research Scholar Award and selection as an EECS Rising Star, as well as multiple Best Paper nominations. It has been supported by NIH, Google, the National Center for Advancing Translational Sciences, and institutional grants. I serve as an Associate Editor for ACM Transactions on Computing for Healthcare and contribute to the broader research community through editorial leadership, service as an Area Chair for conferences such as ACL, EMNLP, and NeurIPS, and the organization of international workshops on reliable large language models and AI in healthcare.

Mentoring is central to my work. I enjoy supporting students as they develop ambitious interdisciplinary research agendas and build careers in academia and industry. Students I have mentored have gone on to positions at organizations including Google, Abridge, Salesforce, MIT, Caltech, Columbia, and Harvard. I welcome collaborations from prospective students and colleagues around human-centered NLP, uncertainty-aware AI, health communication, and the responsible design and evaluation of AI systems.

Education

Ph.D. University of Virginia

M.Sc. University of Virginia

B.Sc. Bangladesh University of Engineering and Technology

Taught Courses

COSC 89/189: Human-centered Large Language Models

COSC 74/274: Machine Learning and Statistical Data Analysis

CS 72/Ling 48: Accelerated Computational Linguistics

COSC 89/189:Human-centric Language Models

COSC 89/189: Computational Healthcare

Publications

Joseph Gatto, Omar Sharif, Parker Seegmiller, Sarah M. Preum. Large Language Models for Document-Level Event-Argument Data Augmentation for Challenging Role Types. The 63rd Annual Meeting of the Association for Computational Linguistics (ACL) 2025. (Nominated for best paper)

Joseph Gatto, Parker Seegmiller, Timothy Burdick, Inas S. Khayal, Sarah DeLozier, Sarah M. Preum. Follow-up Question Generation For Enhanced Patient-Provider Conversations. The 63rd Annual Meeting of the Association for Computational Linguistics (ACL) 2025.

Parker Seegmiller, Joseph Gatto, Sarah Greer, Ganza Belise Isingizwe, Rohan Ray, Timothy Burdick, Sarah M. Preum. How Much Would a Clinician Edit This Draft? Evaluating LLM Alignment for Patient Message Response Drafting. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL), 2026.

Parker Seegmiller, Sarah M. Preum. Measuring the Effects of Natural Shifts in User Prompt Distribution on Large Language Model Performance. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL), 2026.

Sarah M. Preum, Sirajum Munir, Meiyi Ma, Mohammad Yasar, David J. Stone, Ronald D. Williams, Homa Alemzadeh, and John Stankovic. A Review of Cognitive Assistants for Healthcare: Trends, Prospects, and Future Directions. ACM Computing Surveys, 2021.

Joseph Gatto, Omar Sharif, Sarah M. Preum. Chain-of-Thought Embeddings for Stance Detection on Social Media. Proceedings of the International Conference on Empirical Methods in Natural Language Processing (EMNLP), 2023.

Madhusudan Basak, Omar Sharif, Sarah E. Lord, Jacob T. Borodovsky, Lisa A. Marsch, Sandra A. Springer, Edward Nunes, Charlie D. Brackett, Luke J. Archibald, Sarah M. Preum. Information Needs for Opioid Use Disorder Treatment Using Buprenorphine Product: Qualitative Analysis of Suboxone-Focused Reddit Data. Journal of Medical Internet Research (JMIR), 2025

Parker Seegmiller, Sarah M. Preum. Statistical Depth for Ranking and Characterizing Transformer-Based Text Embeddings. Proceedings of the International Conference on Empirical Methods in Natural Language Processing (EMNLP), 2023.

Omar Sharif, Joseph Gatto, Madhusudan Basak, Sarah M. Preum. Explicit, Implicit, and Scattered: Revisiting Event Extraction to Capture Complex Arguments. Proceedings of the International Conference on Empirical Methods in Natural Language Processing (EMNLP), 2024.

Omar Sharif, Madhusudan Basak, Tanzia Parvin, Ava Scharfstein, Alphonso Bradham, Jacob Borodovsky, Sarah Lord, Sarah M. Preum. Characterizing Information Seeking Events in Health-Related Social Discourse. The 38th Annual AAAI Conference on Artificial Intelligence (AAAI), 2024.

Bushra Hossain, Sarah M. Preum, Md Fazle Rabbi, Rifat Ara, Mohammed Eunus Ali. Subreddit to Symptomatology: A Lexicon-based Approach to Extract Symptoms of Complex Conditions from Online Discourse. Journal of Medical Internet Research (JMIR), 2025.

Will Romano, Omar Sharif, Madhusudan Basak, Joseph Gatto, and Sarah M. Preum. Theme-driven Keyphrase Extraction to Analyze Social Media Discourse. Proceedings of the AAAI Conference on Web and Social Media (ICWSM). 2024

Speaking Engagements

When Ground Truth Is a Distribution: New Tasks and Evaluation Paradigms for Asynchronous, High-Stakes Communication. Colloquium at Computational Health Informatics Program (CHIP), Spring 2026, Harvard University.

NLP Seminar, Department of Computer Science, Spring 2026, Columbia University.

Language Technology Institute Colloquium, Spring 2026, Carnegie Mellon University.

From Cognitive Burden to Collaborative Intelligence: Socio-Technical AI for Clinical Communication. Human Computer Interaction Institute Colloquium, Spring 2026, Carnegie Mellon University.

Center for Technology and Behavioral Health Seminar (CTBH) Series, Fall 2025.

SYNERGY Translational and Learning Health System Science Collaborative Seminar Series, Fall 2025, Dartmouth Clinical and Translational Science Institute.

Network Science Institute Colloquium, Department of Computer Science, Fall 2024, Northeastern University.

Unveiling the Journey: Analyzing Information Seeking Events in Online Recovery Discourse. Center for Technology and Behavioral Health Seminar Series, Fall 2023.

Intelligent Assistants for Improved Health: Teaching Machines to Understand Data. Biomedical Data Science Grand Rounds, Geisel School of Medicine, Winter 2022, Dartmouth College.

Preparation for Research through Immersion, Skills, and Mentorship (PRISM) seminar, Department of Computer Science, Spring 2021, University of Toronto

Information Extraction & Fusion for Improving Personal & Public Health Safety, Senseable City Lab Seminar, Spring 2020, Massachusetts Institute of Technology (MIT).

Human-centered Artificial Intelligence Center Seminar, Spring 2020, Stanford University.

Future Directions for Clinical Decision Support, Clinical Decision Support Workshop, Fall 2020, Office of the National Coordinator for Health Information Technology.

Towards a cognitive assistant for emergency response, Public Safety Broadband Stakeholder Meeting by National Institute of Standards and Technology (NIST)

Works in Progress

Selected Works & Activities