Critical Data and Machine Learning Studies
Description: In this course, we will investigate the complex ways that such data and machine learning (ML) technologies are shaped by human values and power structures, and in turn, how these technologies shape society. We will work to understand these systems as sociotechnical — thinking beyond a particular technical artifact to explore how problems are defined and deemed important, how development practices and categorization schemes become default, and how the impact of a system can differ widely across communities.
Together, we will develop both critical and imaginative perspectives on data and machine learning systems. A critical perspective will help you analyze technology in the context of power structures and social context; an imaginative perspective will help you use this awareness to envision and build alternative and more just futures. In your future endeavors as technologists and researchers, you can draw on these analytical frameworks to center diverse lived experiences and collective liberation throughout your work.
The questions we will explore in this class are fundamentally interdisciplinary, and this will be reflected in the readings. Most weeks will include 1-2 readings (sometimes optional) that touch on foundational social theory, exploring concepts such as situated knowledge, intersectionality, decoloniality, and positionality. The other readings will focus on recent research in computer science that bridges these concepts with modern ML & AI technology and practice.
Advanced Natural Language Processing, Teaching Assistant
Graduate teaching assistant for Prof. Jacob Andreas. Developed and taught a new course module on dataset design and its ethical implications with a hands-on problem set. This material received positive feedback from ~250 students, has continued to be taught in subsequent years, made widely available on MIT OpenCourseWare, and adapted for different classes.
ML Tidbits
Co-director of ML Tidbits, an educational non-profit intended to empower the public to understand and discuss machine learning concepts and their societal effects. Wrote, illustrated and published engaging short videos on YouTube with quantitative pedagogical benefits.
MIT AI Ethics Reading Group
Co-founded and led an MIT-wide initiative to connect community members on topics around AI Ethics and facilitate interdisciplinary conversations. Organized bi-weekly meetings, curated readings, and facilitated discussions.
Intro to Deep Learning
Created and led an extensive introduction to the field of deep learning, covering applications to machine translation, image recognition, game playing, image generation and more. Counted as a for-credit course at MIT. Included hands-on labs in TensorFlow and peer brainstorming sessions. 250+ students attended. RNN lecture video here; slides here.