About Me I am Sean Kent

Hi! I am Sean Kent, a Statistics Ph.D. candidate in my fifth year at the University of Wisconsin – Madison. I research weakly-supervised machine learning with outcomes that have ordinal labels under Professor Menggang Yu.

Here, I highlight my current research, statistical software, and side projects. The side projects—often featured on my blog—have recently focused on the analysis of casino games and the improvement of data-science visualizations.

My Insights Observations and Shareable Knowledge



  • Kent S, Yu M, Liu Y (2021+). Multiple Instance Learning from Distributional Instances, submitted.


  • Fiore M, Smith S, Adsit R, Bolt D, Conner K, Bernstein S, et al. (2021+). The First 15 Months of the COVID Pandemic: Changes in Mortality, Intubation and ICU Rates, and Number of In-Patient Days Amongst 139,952 COVID Patients Hospitalized at 21 United States Health Systems, submitted.
  • Grady K, Cameron S, Kent S, Barnes Heller H, Barry M (2021). Effect of an intervention to increase exercise on sleep and seizure frequency in dogs with idiopathic epilepsy receiving anti-epileptics, accepted in Journal of Small Animal Practice.
  • Barry M, Cameron S, Kent S, Barnes Heller H, Grady K (2020). Daytime and nocturnal activity in treated dogs with idiopathic epilepsy compared to matched unaffected controls, Journal of Veterinary Internal Medicine.


Craps: One Surprising Feature of Dark Side Betting

October 24, 2021

I’ll be honest, I don’t think that I’ve ever played the dark side in craps. However, a user of my craps simulator, Eric Hoffert, found some head-scratching results as he compared the win percentages for some dark and light side strategies. As we will see, the chance of walking away from the table is higher than 50% under some dark side strategies. This phenomenon has a few downsides, of course, but this is surprisingly different from most pass-line strategies. In this post, we’ll dive deep into this surprising feature of the don’t pass strategy and discover more unexpected results along the way. [Read More...]

5 Systems to Try at the Craps Table

February 22, 2021

I have often thought about testing some of my craps theories. The folks at Best Craps Strategy provided the impetus to look deeper. On their site, they have a page on craps systems: detailed strategies usually involving multiple bets at different times. Since it is tricky to know how the combination of bets will affect wins and losses, you can’t just look up the odds for these systems. My simulator offers an opportunity for a detailed and rigorous analysis. [Read More...]

Summer of the Pandemic: Visualizing COVID-19 in Wisconsin

October 5, 2020

In March—which feels like forever ago—the spread of COVID-19 in the U.S. was undeniable and many were left wondering what to do. At that point, there was a lot of information circulating about COVID-19, but not a ton of centralized resources to understand the broad patterns. I wasn’t sure where to start, but I decided I wanted to help however I could. [Read More...]

Turning a Bad Visualization into a Good One with Beer Data

September 29, 2019

This blog post is all about personal growth. Last year, I was playing around with some data I scraped from BeerAdvocate’s website. In particular, I was looking at their Top 250 Rated Beers list, which contains information on the beers that BeerAdvocate users rate highest. From that data, I created the visualization in the following tweet [Read More...]

See all blog posts

Highlighted Projects

Software Packages

  • mildsvm is an R package for training classifiers on multiple-instance learning (MIL) data—sometimes called weakly-supervised data. It supports binary and ordinal predictions based on classical and state-of-the-art methodological research. Researchers should use the MIL setting when instances in a data set have hidden labels, but groups of instances have a label.
  • rsmatch is an R package designed for causal matching in time-varying observational studies. It implements two popular methodologies: Balanced Risk Set Matching (Li, Propert, and Rosenbaum, JASA, 2001) and Propensity Score Matching with Time-Dependent Covariates (Lu, Biometrics, 2005).
  • crapssim is a python based simulator for the casino game craps. It can flexibly simulate many scenarios because it takes an object-oriented approach. I've used crapssim to fairly compare the plausible outcomes of different strategies across millions of rolls.


Education & ExperienceMore about my past

2017 – Present

University of Wisconsin – Madison

I'm currently a fifth-year Ph.D. candidate in the Department of Statistics. My research explores new theoretical extensions to multiple-instance learning with ordinal labels under Professor Menggang Yu. In May 2020, I completed my M.S. in Statistics and received a letter of merit for outstanding performance on the Master's Exam on statistical consulting. I'm actively involved in statistical collaborations with the Health Innovation Program, the School of Veterinary Medicine, and the AFI Data Science Institute.

2013 – 2017

University of Notre Dame

I completed my B.S. in Mathematics, with a minor in Actuarial Science from Notre Dame in 2017. I was a member of the Glynn Family Honors Program, where each year, 100 students take general education courses that emphasize discussion and writing from top professors. In my spare time, I was an active participant in my residence hall community— running its pizzeria, Keough Kitch, leading a team of 17 as the 2016 Welcome Weekend Captain, and serving as its representative in the Student Government Senate.

Research Experience
May 2019 – Present

Ph.D. Dissertation Research

[Madison, WI] Under the supervision of Professor Menggang Yu, I improved the state-of-the-art methodology in a weakly-supervised subset of machine learning, known as ordinal regression (OR) for multiple-instance learning (MIL). To complement this research, I developed an R package for MIL algorithms, mildsvm, implementing popular approaches from the literature and our proposed methods for distributional instance and ordinal regression. We applied our methodology to breast-cancer imaging, tissue micro-array, image recognition, and text document classification.

Work experience
Summer 2021

Biostatistics Intern | Pfizer

[Remote, Pearl River, NY] Working with the vaccine assay team, I resolved an unsettled FDA question and improved statistical power by researching top goodness-of-fit measures in probit regression for vaccine in-vivo potency assay. In addition, I saved ~30 hours of statistician time per quarter by developing an interactive app, in R Shiny, to verify the lab data’s accuracy, and I automated an assay robustness analysis—including figures, tables, and data diagnostics—by building an interactive app, in R Shiny.

Summer 2018

Statistical Modeling Intern | Equifax for Business

[Skokie, IL] Continuing on my previous experience, I modeled the compensation of small-business owners by estimating lognormal distribution parameters from geographical information and PayNet business factors using R and SQL.

Summer 2017

Statistical Modeling Intern | Equifax for Business

[Skokie, IL] I investigated the long-term performance of a probability-of-default model across 4 key metrics and 100’s of category combinations by developing a nearly-automated procedure in R and SQL. I also demonstrated an order-of-magnitude speed up to a separate model-training process—while maintaining in-production accuracy—by implementing neural networks and boosting algorithms in Python.

2015 – 2017

Owner/Operator | Keough Kitch

[Notre Dame, IN] My friend and I took ownership over a small pizza restaurant in Notre Dame's Keough Hall for 2 years. We implemented a data-driven approach to menu development and marketing, resulting in sales above all records and a strong presence in the residence hall community that was missing in previous years. I was directly responsible for managing 8 employees, recording and analyzing sales data, and executing marketing strategies.

Summer 2015

Analytics Intern | Equifax for Business

[Skokie, IL] In my first professional internship, I helped predict the probability-of-default for any US country and industry combination by developing a Bayesian hierarchal model in R. I also discovered insights by querying 23 million small-business loans in SQL and helped automate some analyses in Excel.

My Skills My special expertise

Find me
Contact Info

Copyright 2017 Sean Kent All Rights Reserved | Design By W3layouts