Where To Start: Interventions Into Data Bias Using Wikipedia
Where To Start: Interventions Into Data Bias Using Wikipedia
Submitter: Sofia Stancioff, Concordia University, sstancioff@gmail.com
Abstract:
Large transformer model based NLP's require equally large training sets, and Wikipedia has become an increasingly popular source of data. This session will discuss the variety of biases that are baked into our language and cultural narratives which become amplified by algorithms. Using a Wikipedia edit-a-thon as a site of engagement, histories of discrimination can be traced from the past through our present and into possible futures. The edit-a-thon will allow participants to pair discourse with action, grounding the discussion in concrete examples and possibilities for immediate intervention.
Areas of STS Scholarship: Big Data, AI and Machine Learning, Social Movements and STS, Feminist STS
Authors/Participants: Sofia Stancioff, Concordia University
