Sensing without Sensors

  • Jonny Canny | UC Berkeley

This talk is about tapping the enormous amount of sensor data that we generate about ourselves every day. We’re not talking about traditional sensors, but about mobile devices, mice, keyboards, microphones, cameras etc. that we interact with all the time. Our first explorations of “sensorless” sensing used voice data captured on mobile phones to measure mood and depression. Our most recent work uses ordinary mouse movement to measure stress. From raw mouse movement, we build models of muscle stiffness and find that when people are stressed, they become “tense” in a way that we can reliably measure. We are exploring generalizations of the muscle tension approach to mobile devices.

The second part of the talk will explore the opportunities and risks of rich behavioral data from invisible sensors and social media. These data blur the boundaries between “quantitative” and “qualitative” analysis and suggest the development of new practices that straddle those disciplinary boundaries.

Speaker Details

John Canny is a professor in computer science at UC Berkeley, and holds the Paul and Stacy Jacobs distinguished Professorship in Engineering. His recent work is in health technology, educational technology, and “Designing with data”. His work marries human-computer interaction and social science with signal processing and machine learning. He has designed and deployed production behavioral modeling systems for several Fortune-500 companies, and is a cofounder of Mark Logic Corp.

Series: Microsoft Research Talks