EquiPy
A Python package for post-processing fairness: it takes a trained model and returns a calibrated, fairer one, for a single sensitive attribute or several at once.
Software · Research programmes · Experiments
Open-source packages, research collaborations and side experiments. Some are peer-reviewed, some are just curiosity with a version number.
A Python package for post-processing fairness: it takes a trained model and returns a calibrated, fairer one, for a single sensitive attribute or several at once.
Spatiotemporal modelling and NLP over news and social posts to detect outbreak signals early, as part of the Mathematics for Public Health initiative.
Identifying Environmental, Social and Governance concepts in the disclosures of Canadian companies, with Algora Lab (UdeM / Mila).
Roughly a hundred notebooks, slide decks and datasets covering Python, numerical methods, Bayesian ML, NLP and actuarial data science, all free to reuse.
More experiments live on GitHub, including the notebooks behind the courses on the teaching page.