I am a PhD student at University College London in the Department of Statistical Science, supervised by Jeremias Knoblauch and Edwin Fong. Before this, I worked as a research assistant at Aalto University supervised by Aki Vehtari.

My research interests include Bayesian probability theory and methodology, predictive model evaluation and comparison, and prior specification. I have previously held data science positions at 7Bridges Ltd., GOGOX Ltd., and VoiceIQ Ltd.

Please reach out for collaboration on future projects!

Bio

2023 -
PhD candidate in Statistical Science UCL Supervised by Jeremias Knoblauch and Edwin Fong
2021 - 2023
MSc Machine Learning, Data Science, and Artificial Intelligence Aalto University Thesis on "Efficient estimation of selection-induced bias in Bayesian model selection," supervised by Aki Vehtari
Research Assistant in agile probabilistic AI Aalto University Probabilistic Machine Learning (PML) group
2019 - 2020
Year abroad studying Mathematics and Computer Science Université de Bordeaux
2017 - 2021
BSc (Hons) Mathematics with a Modern Language (French) University of Manchester Thesis on "Graphical models for time series analysis," supervised by Jingsong Yuan and Korbinian Strimmer

Selected research

  1. Yann McLatchie, Edwin Fong, David T. Frazier, and Jeremias Knoblauch (2024). "Predictive performance of power posteriors." arXiv. arXiv code
  2. David Kohns, Noa Kallionen, Yann McLatchie, and Aki Vehtari (2024). "The ARR2 prior: flexible predictive prior definition for Bayesian auto-regressions." arXiv. arXiv code
  3. Yann McLatchie and Aki Vehtari (2024). "Efficient estimation and correction of selection-induced bias with order statistics." Statistics and Computing. arXiv code blog
  4. Yann McLatchie, Sölvi Rögnvaldsson, Frank Weber, and Aki Vehtari (2024). "Advances in projection predictive inference." Statistical Science (to appear). arXiv code blog

Software

I help develop the following software:
Kulprit
Kullback-Leibler projections for Bayesian model selection author. code
LOO
Approximate leave-one-out cross-validation (LOO-CV) and Pareto smoothed importance sampling (PSIS) contributor. code
Projpred
Projection predictive feature selection contributor. code
Bambi
Bayesian model-building interface in Python contributor. code