About

I am a PhD candidate in statistics at UBC, where I also completed my MSc. My supervisors are Trevor Campbell and Ben Bloem-Reddy. I did my BSc in applied math at ITAM under the supervision of Ernesto Barrios.

Before coming to UBC, I was a business analytics coordinator at Pfizer. I was also part of the quick counts team during Mexico’s 2018 presidential election, working as a research assistant for Manuel Mendoza. Check this webpage if you’re interested.

In my free time, I enjoy running and road cycling around Vancouver. I also like reading; recently, I’m reading a lot of contemporary female Latin American writers—thanks mainly to recommendations from my wife Sarah.

Research interests

I am mostly interested in developing algorithms for Bayesian inference over discrete variables. This includes general-purpose methods for this setting but also specialized algorithms for specific types of discrete structures, such as directed acyclic graphs in causal discovery.

I am also interested in statistical education, and particularly in finding ways to improve and expand the teaching of Bayesian statistics in the undergraduate curricula.

News

• [August 2024] I successfully defended my thesis proposal, which means that I’m officially a PhD candidate!

• [May 2024] Our paper on a new fully-online methodology for assessing educational resources in development has been accepted to the International Conference of the Royal Statistical Society!

• [Jan 2024] Our paper on learning discrete distributions with MCMC-inspired variational flows has been accepted at AISTATS!

• [Sept 2022] Our paper on using structured and interactive resources to teach Bayesian inference has been accepted at ICOTS!

Publications

Peer-reviewed

• Mixed variational flows for discrete variables
  Gian Carlo Diluvi, Benjamin Bloem-Reddy, and Trevor Campbell
  International Conference on Artificial Intelligence and Statistics, 2024
  copy bibtex   arXiv   AISTATS   code   slides   poster

• Structured, interactive resources for teaching Bayesian inference
  Gian Carlo Diluvi, Bruce Dunham, Nancy Heckman, Melissa Lee, Rodolfo Lourenzutti
  International Conference on Teaching Statistics, 2022
  copy bibtex   pdf   slides

Workshops and conferences

• Using online student focus groups in the development of new educational resources
  Gian Carlo Diluvi*, Sonja Isberg*, Bruce Dunham, Nancy Heckman, Melissa Lee
  Royal Statistical Society International Conference, 2024
  copy bibtex   arXiv   slides   poster

Theses

• Locally-adaptive boosting variational inference
  Gian Carlo Diluvi
  University of British Columbia, 2021 (MSc thesis)
  copy bibtex   pdf   code   slides

• Bayesian design of experiments for generalized linear models
  Gian Carlo Diluvi
  Instituto Tecnológico Autónomo de México (ITAM), 2018 (BSc thesis)
  copy bibtex   pdf   code

Teaching

Mentorship:

• TA trainer, department of statistics @ UBC, 2021—2024

TAing:

• STAT 450 case studies in statistics, spring 2025 @ UBC
• STAT 450 case studies in statistics, spring 2022 @ UBC
• STAT 302 introduction to probability, spring 2020 @ UBC
• STAT 200 elementary statistics for applications, spring 2020 @ UBC
• STAT 203 statistical methods, fall 2019 @ UBC
• EST 14102 probability II, spring and fall 2016 @ ITAM

Flexible learning: an initiative that brings together open and vetted resources to help promote the teaching and learning of introductory statistics.

Here are the resources that I have developed:

• First Bayes
  beta-binomial dashboard   normal-normal dashboard   code  

• Power and sample size calculations
  dashboard   slides   code   StatSpace

• Effective data visualization with R
  dashboard   code   StatSpace

• Understanding type I errors, type II errors, and p-values
  StatSpace

• Distribution of the p-values under the null hypothesis of a one-sample z-test
  StatSpace

Contact

If you are interested in doing an MSc or a PhD in statistics at UBC, reach out in English o en español. Check my UBC stats webpage for contact info.