Newest news! After a wonderful 3 years at University of Toronto, I'm moving to Montreal to become a core member of Mila, with a tenure-track position at HEC. Tiohtià:ke/Montréal is on unceded Indigenous lands. I recognize the Kanien’kehá:ka Nation as the custodians of the lands and waters in this place I call home. To them I offer my deep thanks, my joy at the beauty and diversity here, and my committment to truth and reconciliation (of which this statement is one small step).
"Mila, awesome, congrats! Uh, isn't HEC like, a business school? Do you..." trails off anyone who knows me in confusion. Do I business? I can't say that I do, but it's one of the best management schools in the world. I had something of an epiphany about the importance of human resources and management science last year, thanks to the scholars at Makerere AI lab I had the fortune to visit and discuss AI deployment with. And, I'm increasingly jaded about the impact Responsible-/ FATE-/ Safe-/ Climate-/ Etc-AI research can have from within the AI community -- we have largely handed over decision-making power w.r.t. AI to business management, so I'm excited about working with other experts at HEC to develop a new program in Responsible AI Management (I'm also excited about ways to transfer power and agency back to communities - reach out if you'd like to talk about that!). Every new generation of students I've been lucky to work with has an increased awareness of the socio-environmental impacts of AI/Business, and our responsibility to each other and to the planet to do better. I see a lot of appetite for a program like this, and I'm excited for graduates of this program to be the leaders of tomorrow.
Of course that's at least a few-year project; in the meantime I'm still primarily running a research lab (see my Research Interests below), which finally has a name and website! ERRATA*™ Lab: Ecological Risk & Responsible AI; Theory & Applications *learning from mistakes. I'm not jazzed about the whole idea of branding, but I *am* jazzed about clearer communication, and I think this achieves that - summarizing my research interests, making it obvious the lab is an entity much larger than me, talking up my awesome students/collaborators and the things we do together. I am recruiting for Fall 2025! see lab website for details/advice/instructions.
I'm restarting my open mentorship hours -- these are flexible/by appointment, with priority given to those currently minoritized in ML. BIPOC, queer, global south, disabled, first-gen scholars, category-averse lovely weirdos, you're all warmly welcomed!
If you have contacted me in the last year or so and did not get a reply/autoresponse, I'm sorry! I had long covid and some other health/personal turmoil, and took a mental health break from all social media (highly recommend, not going back any time soon). Also outlook marks a LOT of valid emails as spam. Please reach out again, using my Mila address :).
(This is the latest entry; click here for full blog. Send me an email titled "Subscribe to blog" to get my blog posts in your inbox.)
Research Interests
Having entered AI in the era where we we definitely didn't call it that, I've seen us go from exciting curiosity-led empirical science to Silicon Valley Corporatopia's Exploitamatic McProduct. It's painfully obvious to me that we are doing AI really wrong, and I don't mean the architecture or optimizer or presence/absence of reasoning module. My most general career goal currently is to explore fundamentally different ways of creating and understanding AI in order to make the future suck less. I'm looking forward to joining the Abundant Intelligences initiative, starting an open makerspace at Mila, establishing art-tech collaborations with SAT, and writing more fiction. Some things I've been consistently interested in and influenced by over the last few years are: Truth and reconciliation, chaotic dynamical systems, Black feminism, ecological futurism, risk, transition design, queer and disability justice, curation, community agency.
In terms of AI, I'm broadly interested in studying “what goes into” deep models - not only data, but the learning environment including task design/specification, loss function, and regularization; as well as the wider societal context of deployment including privacy considerations, trends and incentives, norms, and human biases. I'm concerned and passionate about AI ethics, safety, and the application of ML to environmental management, health, and social welfare.
My goal in research is to contribute understanding and techniques 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. My recent research has 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. Please contact me if you're interested in collaborations in these areas. For more detail on ongoing projects, hiring/recruiting, etc., please see my ERRATA*™ Lab Website.
Biography
I started post-secondary education in biology with a focus on health and neuropsychology, but transitioned to a concentration in ecology as I learned about invertebrates, microbiology, and all the amazing interactions of life around us. Analyzing results for my honour's research in the community ecology of bioremediation, I was introduced to programming for the first time, and quickly realized I wanted to use machine learning to understand and model complex ecosystems. I recieved an NSERC scholarship to particpate in a large-scale research project on climate change, and later participated in a number of coding projects and discovered neural networks.
I began an MSc in computer science with Layachi Bentabet, studying biological realism in deep networks. During this time I was awarded a MITACS scholarship to be a machine learning research intern at iPerceptions, exploring semi-supervised learning in predictive models (aka clickstream prediction aka discovering firsthand how right jeff hammerbacher was/is)
In November 2015 I completed my MSc, and in January 2016 began a PhD at Mila (then Lisa), a world-leading academic research institute in Montreal for AI and deep learning, where I was an NSERC- and IVADO-awarded scholar with Christopher Pal. I did a lot of different projects aimed at understanding how deep learning systems perform in "real world" conditions, and my thesis reviews factors that affect generalization in deep learning. I was involved in creating and structuring the Montreal Declaration on Responsible AI. I also helped start the Mila Lab Representatives, became a managing editor at the Journal of Machine Learning Research (JMLR), the top scholarly journal in machine learning, and was a co-founding member of Climate Change AI (CCAI), an organization which catalyzes impactful work applying machine learning to problems of climate change.
Beginning June 2021 I accepted a tenure-track position at the University of Toronto, in the Faculty of Information, where I was an affiliate of Vector, co-founding member of the Toronto Climate Observatory and honoured to be a fellow of the illustriously multidisciplinary Schwartz Reisman Institute for Society and Technology. Information is a unique and curious "field"; I will forever be grateful that I began my career in a place so deeply embodying the idosyncratic and very human pursuit of knowledge and understanding.
CV
My CV can be found here.
Papers
(* denotes equal contribution)
Despite my best efforts to keep this updated, and desire to have better tools for research tracking and credit assignment, my scholar profile remains most reliable for my latest research.
Predicting infectiousness for proactive contact tracing. Yoshua Bengio*, Prateek Gupta*, Tegan Maharaj*, Martin Weiss*, Nasim Rahaman*, Tristan Deleu, Eilif Muller, Meng Qu, Victor Schmidt, Pierre-Luc St-Charles, Hannah Alsdurf, Olexa Bilanuik, David Buckeridge, Gáetan Marceau Caron, Pierre-Luc Carrier, Joumana Ghosn, Satya Ortiz-Gagne, Chris Pal, Irina Rish, Bernhard Schölkopf, Abhinav Sharma, Jian Tang, Andrew Williams ICLR 2021 (oral). [pdf]
COVI-AgentSim: An agent-based model for evaluating methods of digital contact tracing. Prateek Gupta*, Tegan Maharaj*, Martin Weiss*, Nasim Rahaman*, Hannah Alsdurf, Abhinav Sharma, Nanor Minoyan, Soren Harnois-Leblanc, Victor Schmidt, Pierre-Luc St Charles, Tristan Deleu, Andrew Williams, Akshay Patel, Meng Qu, Olexa Bilaniuk, Gaétan Marceau Caron, Pierre Luc Carrier, Satya Ortiz-Gagné, Marc-Andre Rousseau, David Buckeridge, Joumana Ghosn, Yang Zhang, Bernhard Schölkopf, Jian Tang, Irina Rish, Christopher Pal, Joanna Merckx, Eilif B Muller, Yoshua Bengio. [pdf]
Toward trustworthy AI development: Mechanisms for supporting verifiable claims. Miles Brundage*, Shahar Avin*, Jasmine Wang*, Haydn Belfield*, Gretchen Krueger*, Gillian Hadfield, Heidy Khlaaf, Jingying Yang, Helen Toner, Ruth Fong, Tegan Maharaj, Pang Wei Koh, Sara Hooker, Jade Leung, Andrew Trask, Emma Bluemke, Jonathan Lebensold, Cullen O'Keefe, Mark Koren, Théo Ryffel, JB Rubinovitz, Tamay Besiroglu, Federica Carugati, Jack Clark, Peter Eckersley, Sarah de Haas, Maritza Johnson, Ben Laurie, Alex Ingerman, Igor Krawczuk, Amanda Askell, Rosario Cammarota, Andrew Lohn, David Krueger, Charlotte Stix, Peter Henderson, Logan Graham, Carina Prunkl, Bianca Martin, Elizabeth Seger, Noa Zilberman, Seán Ó hÉigeartaigh, Frens Kroeger, Girish Sastry, Rebecca Kagan, Adrian Weller, Brian Tse, Elizabeth Barnes, Allan Dafoe, Paul Scharre, Ariel Herbert-Voss, Martijn Rasser, Shagun Sodhani, Carrick Flynn, Thomas Krendl Gilbert, Lisa Dyer, Saif Khan, Yoshua Bengio, Markus Anderljung. [website] [pdf]
Tools for society [chapter in Tackling Climate Change with Machine Learning]. Tegan Maharaj, Nikola Milojevic-Dupont. Edited by David Rolnick, Priya L. Donti, Lynn H. Kaack, Kelly Kochanski 2019. [website] [pdf]
Hidden incentives for self-induced distributional shift. David Krueger, Tegan Maharaj, Shane Legg, Jan Leike SafeML@ICLR2019. [pdf]
Memorization in recurrent neural networks. Tegan Maharaj, David Krueger, Tim Cooijmans PADL@ICML2017. [pdf]
Reserve output units for deep open set learning. David Krueger, Tegan Maharaj COSL@CVPR2017. [pdf]
A closer look at memorization in deep networks Devansh Arpit*, Stanislav Jastrzebski*, Nicolas Ballas*, David Krueger*, Emmanuel Bengio, Max Kanwal, Tegan Maharaj, Asja Fischer, Aaron Courville, Yoshua Bengio, Simon Lacoste-Julien ICML2017. [pdf]
ExtremeWeather: A large-scale climate dataset for semi-supervised detection, localization, and understanding of extreme weather events. Evan Racah, Christopher Beckham, Tegan Maharaj, Prabhat, Samira Kahou, Christopher Pal. NeurIPS2017. [pdf] [code] [dataset]
A dataset and exploration of models for understanding video data through fill-in-the-blank question-answering. Tegan Maharaj, Nicolas Ballas, Anna Rohrbach, Aaron Courville, Christopher Pal. CVPR2017. [pdf]
Suprisal-Driven Zoneout. Kamil Rocki, Tomasz Kornuta, Tegan Maharaj. 2016. [pdf]
Zoneout: Regularizing RNNs by Randomly Preserving Hidden Activations. David Krueger*, Tegan Maharaj*, Janos Kramar, Mohammad Pezeshki, Nicolas Ballas, Nan Rosemary Ke, Anirudh Goyal, Yoshua Bengio, Hugo Larochelle, Aaron Courville, Chris Pal. ICLR2017. [pdf] [poster] [code]
Practical applications of biological realism in artificial neural networks. [Master's thesis]. 2015. [pdf] [code]
Workshops and other contributions
I've co-organized several workshops:
- Rethinking ML Papers at ICLR 2021
- Climate Change: How Can AI Help? at ICML 2019
- LSMDC (Large Scale Movie Description Challenge) workshop at ECCV2016 and ICCV2017
- Joint Workshop on Storytelling with Images and Videos
- Joint Women in Deep Learning workshop at the Deep Learning Summer School 2016
I was a co-founder of the Montreal AI Ethics meetup, and a contributor to SOCML 2017 and 2018, as well as the Montreal Declaration for Responsible AI and the Beneficial AGI Conference.
I've received outstanding reviewer awards at every venue since NeurIPS began that practice in 2017.
Talks and presentations
Woefully outdated, but watch this space! Updates, teaching materials, and other fun stuff coming soon.LSMDC2016 - Fill in the Blank Challenge. Joint 2nd Workshop on Storytelling with Images and Videos (VisStory) at ECCV. 2016/10. [slides]
Zoneout: Regularizing RNNs by randomly preserving hidden activations. Deep Learning Summer School. 2016/08. [slides]
BRAINS (anatomy, structure, function, and evolution). University of Montreal. 2016/06. [slides]
Neuroscience and biology for deep learning. University of Montreal. 2016/04. [slides]
Introducing "neurotransmitters" to an artificial neural network for modular concept learning and more accurate classification. Research week, Bishop's University, Sherbrooke, QC. (1st prize in poster competition) 2014/02.
Intelligent data analysis broadens our understanding of the world (2nd prize in oral competition) 2014/02.
Teaching
I was a TA for the following classes during PhD:
- Deep Learning
- Artificial Intelligence
- Introduction to Machine Learning
During undergrad and master's:
- CSC211 Introduction to Programming
- CSC103 Interactive Web Page Design
- FIN218 Digital Imaging
- PHY101 Introductory Statistics
- BIO349 Invertebrate Zoology
- ESG226 Oceans I
- BIO110 Genetics
- BIO116 Diversity of Life
I also worked as a tutor at the Computer Science Help Centre at the end of my BSc/beginning of MSc, and at the ITS Helpdesk (troubleshooting and tech support) throughout my BSc.
Software
Maybe it's a bit of a stretch to call it "software", but I wasn't sure where else to talk about this website. It's an honest-to-goodness static website, made entirely by me, and now proudly hosted on Neocities. You can make your own honest-to-goodness not-a-corporate-platform-website there too, which I encourage you to do! It's easier than latex. The rest of the links in this section are relics of this website's former instantiation on a server I set up and managed with the help of my good friend Jordan Slaman, an amazing person who is no longer with us. I'm leaving them up, and dedicate this website, in his memory. It took me a long time to touch this website after he died -- he would NOT have approved! Thanks for everything Jordan. I also dedicate this website to my dad, who taught me html in our basement when I was 8, and has run the honest-to-goodness unofficial Naps website http://www.naparima.org/ since '96.
Prediction and generation of sound with LSTMs: end-of-term project for a deep learning course. [website] (research blog) [code] (based heavily on johnarevalo's code in blocks for RNN-char-prediction, modified to take and generate sound)
Real-time image segmentation: an end-of-term project in a computer vision course. A C++ program segments an image. I also created the web front-end. [website] [code]
S.E.A.N.N. (Software Engineering Artificial Neural Network) group project: Draw a digit and a trained neural network will tell you what probability it assigns to that number being [0-9]. [website] [code]