The DSI Distinguished Speaker Series will highlight expert researchers who are applying data, machine learning, and computational systems to a broader scientific discipline.

Guest Speaker

Andrew Schwartz, Assistant Professor, Department of Computer Science, Stony Brook University


March 1, 2021 (2:00 PM – 3:00 PM ET) – Online Event

Hosted By

DSI Postdoctoral Researchers

About the Seminar

Talk The Digital Phenotype: Towards a Human-Centered AI for Saving LivesTitle and Abstract Coming Soon

Abstract: In many ways, your device knows more about you than your closest friend. With artificial intelligence (AI) techniques, could that someday be in your best interest? Currently, assessments in health care are primarily physical or biological in nature. Scalable psychological information is mostly limited to subjective reports or limited questionnaires. For the first time in human history, a substantial portion of our daily language behaviors is digitally recorded. With care taken for privacy and security, this “digital phenotype” could transform psychological assessments and decisions in health care that can ultimately save lives. However, most of our techniques in AI for analyzing language do not model the person behind the words, let alone optimize toward their psychological well-being.

In this talk, I will present a case for the power of the digital phenotype, combined with AI, to assess psychological attributes and future health-related behaviors with often unprecedented performance. I will discuss work using social media to predict a diagnosis of depression, the future trajectory of PTSD symptoms, and future likelihood of relapse in addiction treatment. I will then discuss work towards breaking down the limitations in AI (and the subfield of Natural Language Processing) to model human data as if it was generated by humans: the need for longitudinal techniques, multi-level modeling of language’s social context, as well as translating multi-disciplinary validity norms. Overall, I suggest a path toward overcoming these challenges will ultimately enable a more robust and powerful AI-based digital phenotype, capable of saving lives.

Bio: Andrew Schwartz received his PhD in Computer Science from the University of Central Florida in 2011 with research on acquiring lexical semantic knowledge from the Web. He then joined the University of Pennsylvania where he was a Postdoctoral Research Fellow and later Visiting Assistant Professor in Computer & Information Science. He is Lead Research Scientist for the World Well-Being Project, a multidisciplinary group of Computer Scientists and Psychologists studying physical and psychological well-being based on language in social media.