Will Epperson

I’m a Ph.D. student in the HCII at CMU advised by Dominik Moritz and Adam Perer.

My research interests lie in developing ways to make complex systems easier to use and understand for experts and non-experts. I am interested in how we can build tools to help users interact with their data through understandable Machine Learning, interactive visualizations, and reccomended analysis. I’m also interested in how these techniques can be applied to the medical domain to improve both doctor and patient experiences.

Education

August 2020 - Present Ph.D. in Human Computer Interaction
Carnegie Mellon University
Advisors: Dominik Moritz, Adam Perer
Sample Coursework: Computational Medicine

August 2016 — May 2020 B.S. in Computer Science
Georgia Institute of Technology
GPA: 4.0, Summa Cum Laude, threads in Intelligence and Modeling/Simulation Sample Coursework: Machine Learning, Deep Learning, Computer Vision, Computer Architecture, Algorithms, Computer Simulation, Information Visualization

Publications

FairVis: Visual Analytics for Discovering Intersectional Bias in Machine Learning
Angel Cabrera, Will Epperson, Fred Hohman, Minsuk Kahng, Jamie Morgenstern, Duen Horng (Polo) Chau
FairVis is a Visual Analytics system that allows users to audit their machine learning models for intersectional bias by exploring the model's performance on various user-specified and reccomended subgroups in a dataset.
IEEE Conference on Visual Analytics Science and Technology (VAST). Vancouver, Canada, 2019.
Project Demo PDF Blog Recording Slides Code BibTeX

			
@article{cabrera2019fairvis,
  title={FairVis: Visual Analytics for Discovering Intersectional Bias in Machine Learning},
  author={Cabrera, {\'A}ngel and Epperson, Will, and Hohman, Fred and Kahng, Minsuk and Morgenstern, Jamie and Chau, Duen Horng},
  journal={IEEE Conference on Visual Analytics Science and Technology (VAST)},
  year={2019},
  publisher={IEEE}
  url={https://poloclub.github.io/FairVis/}
}
		

RECAST: Interactive Auditing of Automatic Toxicity Detection Models
Austin P. Wright, Omar Shaikh, Haekyu Park, Will Epperson, Muhammed Ahmed, Stephane Pinel, Diyi Yang, Duen Horng (Polo) Chau
RECAST is an interactive tool that allows users to audit toxicity detection models with their own input text and suggests alternative wordings for detected toxic speech.
The eighth International Workshop of Chinese CHI. 2020.
Project PDF BibTeX

			@article{wright2020recast,
title={RECAST: Interactive Auditing of Automatic Toxicity Detection Models},
author={Austin P. Wright and Omar Shaikh and Haekyu Park and Will Epperson and Muhammed Ahmed and Stephane Pinel and Diyi Yang and Duen Horng (Polo) Chau},
year={2020},
eprint={2001.01819},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
		

Talks

FairVis
May 2019IEEE Visualization Conference

Honors and Awards

2019PURA: President's Undergraduate Research Award
$1500 research grant to continue work on FairVis project

2016Stamps President's Scholarship
Full ride scholarship given to 40 incoming freshman at Georgia Tech

Research Experience

August 2020 - PresentCarnegie Mellon University, Pittsburgh, PA
Graduate Researcher, Data Interaction Group (DIG)
Advisor: Dominik Moritz, Adam Perer
Member of the DIG research group, working on novel data visulizations, ML interpretation techniques, and interactvie data systems.
Relevant Skills: Python, Javascript

January 2019 - May 2020Georgia Institute of Technology, Atlanta, GA
Undergraduate Researcher, Polo Club of Data Science
Advisor: Duen Horng (Polo) Chau
Member of the Polo Club of Data Science working on novel data visulizations to find fairness issues in Machine Learning models
Relevant Skills: Python, Javascript

January 2018 - May 2019Georgia Institute of Technology, Atlanta, GA
Undergraduate Researcher, Automated Algorithm Design
Advisor: Jason Zutty, Greg Rohling
Worked on EMADE algorithm design engine to implement sentiment analysis pipeline to analyze news articles to aid in predicting stock price movements using genetic algorithms. Led project to visualize the genetic algorithm evolution process.
Relevant Skills: Python, Javascript

Industry Experience

Summer 2019Point72 Asset Management, New York, NY
Data Analytics Intern, Market Intelligence Group
Mentor: Trevor Rempel
Worked as Data Scientist in alternative data space to clean, model, and understand large datasets
Relevant Skills: Python, Distributed Computing in Spark

Summer 2018Ultimate Software, Weston, FL
Software Development Intern, Innovation Strategies Team
Mentor: Joseph Cutrono
Designed and developed Slack app to integrate with the UltiPro HR management tool. App published to Slack app store.
Relevant Skills: Typescript, REST API development

Summer 2015The Home Depot, Atlanta, GA
Software Development Intern
Developed web app for tracking candidate progress throughout hiring process for internal HR use.
Relevant Skills: Java, HTML/CSS/Javascript

Teaching

August 2017 - December 2018Undergraduate Teaching Assistant
Georgia Institute of Technology, Atlanta, GA
Intro to Database Systems (CS 4400), Instructor: Monica Sweat
Designed projects, held office hours and graded for relational databases class.

Leadership & Activities

January 2019 - May 2020 Student Ambassador
Georgia Institute of Technology Alumni Association
Serve as official representative of the Institute at events/tours for alumni, prospective students, and special guests.

January 2018 - May 2019 Executive Board Member -- Threads Co-chair
Stamps Scholars National Convention 2019
Executive board member of Stamps Scholars National Convention, a 3-day conference with over 700 student attendees. Responsible for 20-person committee that plans and coordinates the different content threads of the convention.

Sample Projects

December 2018 ICLR’19 Reproducibility Challenge
Implemented the architecture and reproduced results for an ICLR'19 submission using GANs to decorrelate sensitive data, titled Generative Adversarial Models for Learning Private and Fair Representations.

January 2018 - December 2018 Atlanta Crime Map
Led a team as part of Data Science Club at GT to analyze and visualize crime data on and around GT's campus to provide insight into crime frequency and details for GTPD.

February 2018 Citadel Datathon
Created supervised learning system to predict dangerous road areas to help make cities safer at Citadel and Correlation One's Datathon at Georgia Tech. Team placed top 3.

February 2017 - December 2017 RECONSO Research
Developed interface between core state machine and battery systems on Avionics team of student-led cube satellite project. Learned C while working on project.

Skills

Programing Languages: Python (Advanced), Javascript/Typescript (Intermediate), HTML (Intermediate), SQL (Intermediate), Java (Intermediate), C (Basic)
Toolkits, Frameworks, Software: Pytorch, Scikit-learn, Git, VegaLite, D3, Tableau, MacOS, Windows, Linux
Natural Languages: English (Native), Spanish (Advanced)