FFF Seminar: Michael Muller (IBM Research)

Abstract:

HCI engages with data science through many topics and themes. Researchers have addressed problems with potentially biased data sampling, arguing that bad data can cause innocent software to produce bad outcomes. But what if our software is not so innocent? What sources of bias do we introduce into our data? Based in feminisms and critical computing, we analyze forgetting practices in data science work practices of planning, finding, cleaning, engineering, curating, and labeling the data of data science. We focus on forms of unknowing, in which a powerful party erases knowledge of past wrongs and harms. Forgetting practices can weaken the integrity of data and outcomes. We contribute: (1) a taxonomy of data silences in data work in HCI and data science, which we use to analyze how data workers forget, erase, and unknow aspects of data; (2) a detailed analysis of forgetting practices in machine learning; and (3) a comparison of these practices with other colonial actions.

Bio:

Michael Muller works as a Research Scientist at IBM Research, on the traditional, contemporary, and unceded lands of the Wampanoag and Massachusett peoples (known to colonists as Cambridge MA US). He conducts critical studies of human work in data science, with earlier research in human collaboration and participatory design practices. Michael pursues social justice work through feminist-informed and Indigeneity-informed analyses. Michael has been recognized by ACM as a Distinguished Scientist, and by SIGCHI as a member of the CHI Academy.