Statistical Learning & AI
Building multimodal foundation models, normalizing flows, and AI/LLM systems to uncover structure across massive, noisy, heterogeneous astronomical surveys.
Astrostatistics · University of Toronto
Joshua S. Speagle沈 佳 士
Combining astronomy, statistics, and AI to understand how stars and galaxies form, behave, and evolve.
About Me
I am an Assistant Professor of Astrostatistics jointly appointed between the David A. Dunlap Department of Astronomy & Astrophysics and the Department of Statistical Sciences at the University of Toronto. I am also an Associate Member of the Dunlap Institute for Astronomy & Astrophysics as well as a Member of the Data Sciences Institute. For the 2026–2027 academic year, I am also serving as the (Interim) Director of the MScAC – Data Science Concentration in the Department of Statistical Sciences.
My research lies at the intersection of astrostatistics, data science, and artificial intelligence. Together with the Astrostatistics Research Team (ART), which I co-lead with Gwen Eadie, I develop and apply statistical and machine-learning methods to large datasets to answer complex astrophysical questions and improve how we learn about the universe from noisy, messy data.
I also teach undergraduate and graduate courses in astronomy and statistics, serve on committees within the University of Toronto, and remain active in professional societies including the AAS, ASA-AIG, SSC-DSA, and CASCA. I also help manage the Astronomy & Astrophysics Summer Undergraduate Research Program (SURP), the largest program of its kind in Canada.
- j.speagle@utoronto.ca
- Office
- AB 206 (Astronomy & Astrophysics) and UY 9198 (Statistical Sciences)
Statistics, AI & Astronomy
Our group develops statistical inference algorithms, scientific AI foundation models, and open-source software, applying them to large astronomical surveys to study stars, star clusters, interstellar dust, the Milky Way, and distant galaxies.
Interpretability & Insight
Investigating the mathematical and statistical foundations of deep learning and representation spaces so we can calibrate uncertainty and trust scientific conclusions drawn from AI models.
Inference & Computation
Designing dynamic nested sampling, Markov chain Monte Carlo, and simulation-based inference algorithms that make rigorous Bayesian inference and model comparison tractable at survey scale.
Discovery & Understanding
Applying these frameworks to billion-star photometric, astrometric, and spectroscopic surveys to map the Milky Way in 3-D and trace how stars, star clusters, dust, and galaxies form and evolve.
143Papers
19,873Citations
46h-index
85Mentees
A mostly complete list of my publications can be found on:Browse all publications, metrics & figures →
Research Group & Working With Me
Research Group & Opportunitiesastrostatuoft.com
The Astrostatistics Research Team (ART)
Bridging astronomy, statistics, and AI
I co-lead the interdisciplinary Astrostatistics Research Team (ART) at the University of Toronto with its founder, Gwen Eadie. Our group brings together faculty, postdoctoral fellows, graduate students, and undergraduate researchers across Astronomy & Astrophysics, Statistical Sciences, Computer Science, and the Dunlap Institute to work on shared methodological, AI, and astrophysical problems.
Statistical Methodology, AI & Open Software
We develop Bayesian sampling algorithms, simulation-based inference methods, and scientific AI foundation models, releasing open-source packages and data products built directly around questions in observational and theoretical astrophysics.
Cross-Departmental Community
Our group spans Astronomy & Astrophysics, Statistical Sciences, and the Dunlap Institute, with weekly group meetings, reading groups, hack days, and joint co-supervision.
Working With Me
Current group & alumni directory →Collaboration and mentorship are central to how I do research. Whether you are a prospective postdoctoral fellow, graduate student, undergraduate researcher, or visiting collaborator, here are the core values of our group and the primary pathways for joining.
Equity, Diversity, and Inclusion
For centuries, explicit and implicit discrimination have barred individuals from many backgrounds from participating fully and freely in science. As scientists and people, we have a duty to address these injustices to better both our profession and society at large. I endeavor to create a community in which all members feel that their diverse backgrounds can be an integral part of their identity as a scientist rather than something that excludes them from being a valued member of the scientific community.
Open and Accessible Science
I am a strong proponent of open, accessible, and reproducible science. When conducting research, my collaborators and I make an active effort to share our results along with public data products, interactive visualizations, and open-source code that remains public throughout its development and well-documented when released to the wider community.
Postdoctoral Researchers
Gwen Eadie and I are always glad to hear from prospective postdocs interested in joining the Astrostatistics Research Team. Because openings depend on funding, most postdoctoral fellows join through one of the fellowships below. If one matches your timeline and research interests, please email either of us well before the application deadline.
Graduate Students
Faculty at the University of Toronto do not admit graduate students directly; admissions are handled by our departmental committees. The first step is applying to one of the Master's or PhD programs below. If you are interested in working together, feel free to email me before you apply to discuss potential projects.
University of Toronto Undergraduates
I work with University of Toronto undergraduates both during the academic year and over the summer. To keep applications and review equitable, I advertise most undergraduate projects through the university programs below.
Pathways & programs:
Research Opportunity Program (ROP) Work Study Program (WSP) University of Toronto Excellence Award (UTEA) independent study courses · AST, PHY, STAcasual employment
Canadian Undergraduates
Students at other Canadian universities typically join our group for paid summer research terms through one of the competitive summer fellowships below.
International Undergraduates
Several funded exchange programs support international undergraduates for summer or semester research terms at Canadian universities. Eligibility depends on your home institution and country, so please check the program details below before applying.
Visiting Researchers & External Collaborators
I welcome new collaborations as well as short- or long-term research visits to the University of Toronto. If you are interested in visiting or collaborating on a project, please reach out by email at j.speagle@utoronto.ca so we can discuss scope and funding.