Innovations and Applications of Large Language Models (LLM) and Emerging AI Technologies
In a futuristic revival of the "Master Methodologist" Invited Speaker sessions longtime ISTSS Annual Meeting attendees may remember from past conferences, two researchers at the forefront of technology and health shed light on the ways that large language modeling and other cutting-edge developments in artificial intelligence are impacting the study and treatment of traumatic stress.
Save Me A Seat
Large Language Models (LLM) and other Natural Language Processing (NLP) methods have emerged as tools to improve the study and measurement of stress. Through the analysis of unstructured text data (including therapy transcripts, medical records, and social media), researchers can study linguistic markers of stress, promising both predictive and monitoring capabilities of symptoms levels.
Part 1: Applications of NLP Methods and LLMs
The speakers will elucidate fundamental NLP approaches—including lexicon methods, topic modeling, word vectors, and deep learning models—and their significance to stress research by reviewing findings from the literature. Understand potential sources of clinical text data and explore areas of applications, including the study of patient and provider characteristics.
Part 2: State-of-the-Art LLM Architectures and Their Emergent Properties
Learn about the training steps to align an LLM to the medical domain—specifically, the capabilities of a medically aligned LLM (Med-PaLM 2) at assessing the diagnosis and symptoms of stress- and trauma-related conditions. The speakers will discuss the potential to augment LLM inputs with data from wearable devices and medical sensors that enable models to process numerical physiological and behavioral data in a more principled manner—a process known as "grounding." Lastly, explore considerations to take into account when using LLMs as a tool from a clinician or patient prospective.
Matteo Malgaroli, PhD
Assistant Professor, Department of Psychiatry
New York University School of Medicine
Dr. Malgaroli, who received his PhD from Columbia University in 2018, works in digital mental health at the intersection of psychiatry, AI, and technology. His research focuses on AI applications to behavioral health to monitor psychopathology, inform early detection, and provide intervention feedback. Dr. Malgaroli conducts federally funded research on Natural Language Processing to identify linguistic markers and analyze clinical transcripts with Large Language Models. Dr Malgaroli teaches computational methods to psychology graduate students, psychiatry residents, and fellows. He also studies digital health behavioral interventions as an ideal source of large-scale clinical data and to increase access to behavioral interventions.
Daniel McDuff, PhD
Staff Research Scientist, Google
Affiliate Professor, University of Washington
Daniel McDuff completed his PhD at the MIT Media Lab in 2014 and has a BA and Master's Degree from Cambridge University. Previously, Daniel worked at the UK MoD, was Director of Research at MIT Media Lab spin-out Affectiva. His work has received nominations and awards from Popular Science magazine as one of the top inventions in 2011, South-by-South-West Interactive (SXSWi), The Webby Awards, ESOMAR and the Center for Integrated Medicine and Innovative Technology (CIMIT). His projects have been reported in many publications including The Times, the New York Times, The Wall Street Journal, BBC News, New Scientist, Scientific American and Forbes magazine. Daniel was named a 2015 WIRED Innovation Fellow, an ACM Future of Computing Academy member and has spoken at TEDx and SXSW. Daniel has published over 200 peer-reviewed papers on machine learning (NeurIPS, ICLR, ICCV, ECCV, ACM TOG), human-computer interaction (CHI, CSCW, IUI) and biomedical engineering (TBME, EMBC).