Led by PhD student Phil Van-Lane (co-supervised with Gwen Eadie; ApJ, 987, 152), ChronoFlow uses conditional normalizing flows calibrated on open star clusters to infer stellar rotation-age posteriors across the Milky Way, with press coverage in U of T News and Phys.org.
Our lead-author study (Speagle et al. 2024, ApJ, 971, 136) mapping distances, extinctions, and stellar properties for 170 million stars across the Milky Way using brutus was highlighted as an AAS Nova Featured Image.
The second data release of the Dark Energy Camera Plane Survey (Saydjari et al. 2023, ApJS, 264, 28) catalogued 3.32 billion objects across 6.5% of the night sky in the southern Galactic plane, with press coverage in NSF NOIRLab, Space.com, PBS NewsHour, and Salon.
Co-authored with postdoctoral fellow Biprateep Dey and accepted for a Spotlight Talk at the NeurIPS 2026 Machine Learning and the Physical Sciences (ML4PS) workshop, our new study shows that pretrained tabular foundation models match or outperform tuned bespoke photo-z estimators with 40x fewer training galaxies and mitigate spectroscopic selection bias via in-context calibration using our open-source lazy-tfm package.
Postdoctoral fellows Christian Kragh Jespersen (Schmidt AI in Science & CITA National Fellow), Maryum Sayeed (Arts & Science Fellow), and Chloe Cheng (Research Excellence Fellow), along with doctoral students Ian Zhang and Isabelle Huang, join our research group at the University of Toronto.
Led by postdoctoral fellow Tanveer Karim (with Renée Hložek), our new study introduces leave-one-out Bayesian model predictive scoring to isolate which individual observations drive the preference for evolving dark energy, showing that specific DESI BAO redshift bins rather than Type Ia supernovae drive the tension with ΛCDM.
Mairead Heiger successfully defended her PhD in the David A. Dunlap Department of Astronomy & Astrophysics (co-supervised with Ting Li) and joins the University of Virginia as a joint GECO and CosmicAI Postdoctoral Fellow.
Mairead Heiger and Simran Nerval received the department’s Fieldus Award recognizing a high level of research acumen and service. Mairead was honored for her spectroscopic follow-up of the ultra-faint dwarf galaxies Eridanus IV and Centaurus I, her galactic chemical evolution modeling framework, and her leadership as GASA co-chair.
Led by PhD student Mairead Heiger (with Ting Li and collaborators; ApJ, 1007, 139), our new study uses Bayesian chemical evolution modeling of Milky Way dwarf spheroidal galaxies to constrain Type Ia and core-collapse supernova nucleosynthetic yields and delay-time distributions.
Co-led with Catherine Zucker and Andrew Saydjari (ApJ, 992, 39), we inferred distances and extinctions for over 550 million stars in the DECaPS2 survey using brutus to construct the deepest 3-D dust map of the southern Milky Way plane.
Steffani Grondin and Michael Poon received the department’s Jui Lin (Allen) Yen Award for exceptional graduate research achievement, recognizing Steffani’s doctoral thesis on white dwarf binaries, extra-tidal stars, and dynamical ejections in star clusters.
Led by PhD student Phil Van-Lane (co-supervised with Gwen Eadie; ApJ, 987, 152), ChronoFlow uses conditional normalizing flows calibrated on open star clusters to infer stellar rotation-age posteriors across the Milky Way, with press coverage in U of T News and Phys.org.
Led by PhD student Steffani Grondin (ApJ, 976, 207), our team combined Gaia DR3, GALEX ultraviolet photometry, and machine learning to identify 52 post-common-envelope white dwarf–main-sequence binary candidates across 38 open star clusters, covered by U of T Arts & Science, Phys.org, and SciTechDaily.
Sam Berek and Shivan Khullar received the department’s Jui Lin (Allen) Yen Award. Sam was honored for her paper introducing zero-inflated Bayesian hierarchical count models (HERBAL) to model galaxy–globular cluster scaling relations.
Our lead-author study (Speagle et al. 2024, ApJ, 971, 136) mapping distances, extinctions, and stellar properties for 170 million stars across the Milky Way using brutus was highlighted as an AAS Nova Featured Image.
U of T Statistical Sciences profiled undergraduate researchers Khalid Edris and Junbo (Felix) Li, co-mentored with Dunlap Fellow Mike Walmsley, on fine-tuning the Zoobot foundation model to classify galaxy morphologies from the Euclid space telescope.
Awarded the University of Toronto’s Connaught New Researcher Award in support of early-career faculty research in astrostatistics, Bayesian inference, and survey science.
The second data release of the Dark Energy Camera Plane Survey (Saydjari et al. 2023, ApJS, 264, 28) catalogued 3.32 billion objects across 6.5% of the night sky in the southern Galactic plane, with press coverage in NSF NOIRLab, Space.com, PBS NewsHour, and Salon.
Awarded the 2022 John Charles Polanyi Prize in Physics by the Council of Ontario Universities and the Government of Ontario in recognition of early-career research in statistical astronomy and Milky Way cartography.
Using 3-D dust maps and Gaia space astrometry, our Nature study (Zucker et al. 2022) showed that virtually all nearby young star-forming complexes lie on the expanding surface of the Local Bubble, covered by CfA, the Harvard Gazette, STScI, and Smithsonian Magazine.
PressCenter for Astrophysics | Harvard & Smithsonian
Profiled by the Dunlap Institute for Astronomy & Astrophysics on developing the allsky interactive panchromatic sky visualization and mapping 170 million stars and interstellar dust across the Galaxy.
The release paper for dynesty, our open-source Python package for dynamic nested sampling and Bayesian posterior and evidence estimation, appeared in MNRAS (493, 3132) and has now been cited over 2,200 times.
Co-authored the Nature discovery (Alves et al. 2020) of a coherent 9,000-light-year undulation of interconnected star-forming clouds in the solar neighborhood using 3-D dust mapping with brutus, featured by CfA and the Harvard Gazette.
PressCenter for Astrophysics | Harvard & Smithsonian