Spring 2024 Highlights
The Spring 2024 semester was a busy one for the Population Health Data Science (PHDS) program. In alignment with the objectives of our program, we hosted several events this semester, ranging from seminars to interactive workshops. Each event contributed towards our goal of increasing population health data science literacy and knowledge among our growing community!
We began the semester with the fourth installment of the Health Data Science Distinguished Speaker Series, which began in September 2023. On January 24, Dr. Suresh Venkatasubramanian (Professor of Data Science and Computer Science, Brown University) presented a seminar entitled, “The Moral Salience of the Algorithmic Lens.” Dr. Venkatasubramanian spoke about the existence of an algorithmic lens through which we see the world today, how this lens is imbued with our own values and judgements, and why we must understand this in order to design systems that align with our goals and values.
In February, we co-sponsored the Health Data Science Focused Research Program along with the Hariri Institute, the Evans Center and the CTSI. As part of this effort, we co-hosted an informational workshop on Feb 6 in Hiebert Lounge that included faculty lightening talks, brainstorming discussions, and a networking lunch. Two proposals were funded: ‘Enhancing Models for Breast Cancer Risk Prediction and Bias Mitigation through Clinician AI Collaboration’ (Leads: C. Poynton & K. Batmanghelich) and ‘Multimodal Transformer Architectures for Neuropathology Study of Alzheimer’s Disease’ (Leads: X. Zhang & C. Farris).
On February 15th, we continued the Distinguished Speaker series with Dr. Tamara Broderick (Associate Professor of Electrical Engineering & Computer Science, Massachusetts Institute of Technology), whose presentation, “Toward a Taxonomy of Trust in Data Science Methods,” focused on potential pitfalls in probabilistic data analyses and how to mitigate them. Dr. Broderick demonstrated that dropping a fraction of data can substantially change the conclusions of an analysis under certain conditions. She outlined this and other potential concerns with generalization of analytic results, and ways to address these concerns.
On March 7, Dr. Jessica Hullman (Ginni Rometty Associate Professor of Computer Science, Northwestern University) presented the sixth installment in the series, “Data Visualization for Inference: Opportunities & Challenges.” In this seminar, Dr. Hullman shared how current approaches to designing effective visualizations for data-driven inference and decision-making are limited, as well as how we might overcome these limitations by utilizing alternative frameworks for visualizing and supporting reasoning about uncertainty in data and model predictions.
Later in the month, on March 26, participants had the opportunity to further develop their data visualization knowledge and skills during a workshop on Data Visualization Techniques & Tools, led by Dr. Prasad Patil and Dr. Lukas Weber (Boston University School of Public Health). Workshop attendees learned about key principles and concepts for data visualization and practiced creating data visualizations using R in a hands-on collaborative activity.
Our next event on April 25 was the first in a new series, “Ask the Data Science Expert.” The purpose of this series is to provide an interactive opportunity for attendees to learn more about a topic of interest in health data science through discussions with an expert in the area. Dr. Sara Lodi (Associate Professor of Biostatistics, Boston University School of Public Health) was the featured expert for this event, which focused on the topic of target trial emulation. Two investigators, Drs. Jonathan Jay and Rachel Yorlets, each had the opportunity to describe their real-world studies using target trial emulation, followed by an interactive discussion with Dr. Lodi, who provided guidance and input on the studies.
The final event for the semester took place on June 11. This Lunch and Learn event, titled “Why Veridical Data Science for Medical AI?” featured Dr. Bin Yu (Chancellor’s Distinguished Professor, University of California at Berkeley), who spoke about promoting veridical (truthful) data science using the PCS (Predictability, Computability and Stability) framework for mitigating the “dangers” of AI, and ongoing research in the realm of PCS uncertainty quantification.
Thank you to everyone who participated in and supported these events this semester. We hope to continue offering similar events next year, and we hope you can join us to kick off the Fall 2024 semester at the second annual PHDS Poster Session & Reception on Thursday, September 5th from 12-2pm in Hiebert Lounge!
