[seminar]

Speaker: Bryn Loftness

Speaker bio:
Bryn Loftness is a doctoral student and NSF Graduate Research Fellow in the Complex Systems and Data Science program at University of Vermont. She collaborates primarily within the M-Sense Research Group developing digital phenotyping algorithms for early childhood mental health disorders using wearable sensors. She also works remotely with the Harvard Sabeti Lab through the Broad Institute of Harvard and MIT specializing in behavioral phenotyping and digital epidemiology projects. Bryn engages in a wide variety of research across her collaborations, with an overall vision for developing novel computational methods supporting the betterment of community, physical, and mental health across populations.

Talk description: 
Digital contact tracing tools, such as those leveraging Bluetooth or Wi-Fi colocation networks, are increasingly utilized for infectious disease surveillance. However, serious privacy risks accompany their use, and the capture of epidemiologically irrelevant contacts can reduce their reliability. Leveraging a robust multimodal dataset collected at Colorado Mesa University during the 2020-21 academic year, Bryn’s team applied white box statistical tools to produce a sparsified network pruned based on normalized behavioral tendency between individuals. This new method enhances mathematical transparency, improves computational efficiency, supports individual privacy, and shows evidence of methodological generalizability.

WATCH THE RECORDING OF THE SEMINAR BELOW: