Research Assistant Position: Machine Behavior/Machine Psychology
Colin Conwell, Vision Sciences Laboratory, Psychology/FAS
fall 2019


Potentiometrics | Psychoanalyzing the Ghost in the Machine: Automated, artificially intelligent systems (powered by machine learning) are increasingly embedded in almost all aspects of contemporary human life, making it imperative we understand how these systems really behave in circumstances outside the relatively rigid testbeds in which they're typically fashioned. Using a modified set of tools from cognitive neuroscience and classic psychophysics, we've begun to probe more deeply the 'black box' of modern machine learning algorithms, contrasting machine behavior with human behavior in ways that extend beyond superficial benchmarks. We believe these tools are best applied and most thoroughly supplemented with a diverse, interdisciplinary array of perspectives that extend far beyond the engineering inherent to their design, and while programming is a plus, we welcome students from any and all backgrounds. Join this project for a chance to learn more about machine learning (deep neural networks), psychology, experimental design and data analysis. Student Involvement: Student involvement in this project will depend on the student's background, as there a wide variety of tasks involved. All members will be expected to attend a weekly meeting (the 'Potentiometrics' meeting) dedicated to collective brainstorming and collaborative hack-a-thons. We put a heavy premium on students' using their idiosyncratic strengths for the good of the group, leading discussions on topics of personal expertise or tutorials on relevant skills. Students may opt to participate for course credit. To Apply: For more information or to apply, please contact Colin Conwell at conwell@g.harvard.edu. (posted 9/2019)