I am a Sociologist and a PhD candidate in Computer Science at TU Berlin. I work as a researcher at the Weizenbaum Institute. I investigate work practices of data production and focus on power dynamics and their effects on machine learning datasets. Broadly, I am interested in questions of meaning making and symbolic power encoded in training data. My work comprises ethnographic fieldwork with data annotators, collectors, and scientists in different sites around the world.
I am also a mom, an immigrant, and a first-generation academic. My pronouns are she/ella.
Studying Up Machine Learning Data: Why Talk About Bias When We Mean Power?
Milagros Miceli, Julian Posada, and Tianling Yang.
[forthcoming] In Proc. ACM Hum.-Compt. Interact. 2021
ACM digital library // PDF //
Wisdom for the Crowd: Discursive Power in Annotation Instructions for Computer Vision.
Milagros Miceli and Julian Posada.
In T. Gebru, E. Denton, (eds.), CVPR 2021 Workshop Beyond Fairness: Towards a Just, Equitable, and Accountable Computer Vision, June 25, 2021.
Documenting Computer Vision Datasets: An Invitation to Reflexive Data Practices
Milagros Miceli, Tianling Yang, Laurens Naudts, Martin Schuessler, Diana Serbanescu, and Alex Hanna.
In Conference on Fairness, Accountability, and Transparency (FAccT ’21), March 3–10, 2021.
Between Subjectivity and Imposition: Power Dynamics in Data Annotation for Computer Vision.
Milagros Miceli, Martin Schuessler, and Tianling Yang.
In Proc. ACM Hum.-Comput. Interact. 4, CSCW2, Article 115 (October 2020).
Best Paper Award
Biased Priorities, Biased Outcomes: Three Recommendations for Ethics-oriented Data Annotation Practices.
Gunay Kazimzade and Milagros Miceli.
In Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society (AIES ’20). Association for Computing Machinery, New York, NY, USA.
Making Data, Making Reality. Power, Visibility, and the Production of Datasets for ML.
In ACM Celebration of Women in Computing – WomENcourage ’20 (24.-26.09.2020). Baku, Azerbaijan.