ERRATA*™ Lab

About

ERRATA* lab is headed by me, Tegan Maharaj. As of October 2024 it's headquartered at Mila - some members remain at University of Toronto, where I began my career as a professor. We work to understand, critique, and change how AI systems perform in real-world contexts, from learning theory and interpretability to on-the-ground situated impact and scenario planning. Our work often integrates technical and socio-environmental analysis, and many of our members come from "non-traditional" backgrounds (lucky us!). At all levels, we aim to foster an environment that is safe, thoughtful, kind, creative, curious, multidisciplinary, collaborative, activist, non-violent, and pragmatic, where we help one another to learn from mistakes -- ours or those of others :) -- imagine better futures, and do good science. We contribute to the growing science of responsible AI development, while usefully applying AI to high-impact ecological problems including biodiversity, climate change, epidemiology, AI alignment, and ecological impact assessments. Most of our research falls in the following three themes:

• (1) Using deep models for policy analysis and risk mapping
• (2) Designing data or unit test environments to empirically evaluate learning behaviour or simulate deployment of an AI system
• (3) Foundations of anticolonial AI

Projects

Coming soon!

Team

Coming soon!

FAQ

Q: What should I do if I want to talk about doing a Master's or PhD with you?
A: I am recruiting for Fall 2025! Send me an email; attach your CV and a paper/piece of work/project you're proud of. Briefly describe your research interests, why you want to work with me/my lab, and what you think is significant about the work you attached/why you're proud of it. Also start planning for the application process and timeline -- graduate applications are usually due around Nov/Dec!

Q: Will you be working in person at Mila?
A: I, Tegan, will be there 4-5 days a week. Very happy to get coffee or lunch or talk research if you're around. Other students and collaborators live in different places with different schedules, and for the foreseeable future lab meetings will be remote or hybrid.

Q: Wait isn't Mila already a lab, and that's where you are? Why do you have another lab name?
A: "Lab" is kind of an amorphous concept in computer science; it's common practice for a professor to call the set of all students/projects/postdocs they supervise (plus maybe other projects they're significantly involved in that don't have another institutional "home") their "lab". Cynically, you could say this is just for branding/SEO, but I do think it's helpful to give a title to this entity that is larger than just me, but smaller than all of Mila. It's not wrong to call Mila a lab (it's a not-for-profit research institute with a pretty unique org structure), but in a big lab-tent like Mila with so many professors, a lab name can be helpful for understanding the more specific kinds of areas that I and my collaborators are focused on. A lot of other Mila profs name their labs, and a lot don't ¯\_(ツ)_/¯. The main downside, given the highly collaborative environment encouraged at Mila, might be over-profusion of acronyms/logos on posters, but this doesn't seem like a big deal to me, and I don't mind at all if you omit the lab name/logo for clarity/brevity anywhere and just say Mila or Tegan's lab or anything else accurate.