Explainable AI: Bridging the Gap between Software and Customers; Robot Umpires in Baseball: How is New Technology Impacting Higher-Revenue Sports?
Bristow, Carter, School of Engineering and Applied Science, University of Virginia
Rogers, Hannah, EN-Engineering and Society, University of Virginia
Vrugtman, Rosanne, EN-Comp Science Dept, University of Virginia
Last summer, I spent a lot of my time focussing on two things: completing my internship with Capital One and following the Olympics, especially swimming, as I had teammates competing. My internship with Capital One was in their technology department, and I was assigned to a Machine Learning team, where I worked closely with how Machine Learning applies to their Credit Card department. It was in meetings with my managers and mentors that I was able to discuss the use of Machine Learning in other areas of my life. We one day looked at how Machine Learning and image recognition were used in athlete tracking in the Olympics that year. In swimming, for example, the instantaneous speeds of the athletes would flash up on the screen as races finished so it is now easier to tell who was winning in a close race or if someone were speeding up or slowing down. These kinds of conversations inspired me to look more into the realm of technology as it relates to sports. I thought that it would be interesting to take the ethical approach that I would cover in an STS research paper on technology in sports and be able to apply it to the technical skills that I learned through the various projects in my internship.
Upon my initial research of technology in athletics, I discovered that the Robot Umpire in baseball is one of the newest technological advances in the sports world. As an already statistics-heavy and high-revenue sport, I thought it would be interesting to tackle the ethical issues of this new development. Throughout my research process, I learned more about the rich history that baseball has in becoming “America’s pastime” and the desire to keep the sport honest to the way it was originally created as well as the need to appeal to fans and make sure rules are upheld. Using similar methods to Don Norman’s Design of Everyday Things and Tamir and Bar-eli’s piece on the Video Assistant Referee in soccer, I created a sentiment analysis program that would analyze different comments from interviews made about the Robot Umpire and label them on a scale from -1 to 1 which indicates how positive, negative, or neutral they are. Additionally, I performed a verbal analysis of the technology on the components of “Athletic Integrity” to see if it upheld integrity as well as received positive feedback from the public. Overall, I discovered a very neutral position on the technology, as the verbal analysis was split on support of athletic integrity and the sentiment analysis produced an average of .195, but I did confirm that technology is making its way into sports in a way that needs to be addressed more often. An analysis similar to the one that I used should be used to evaluate new technologies before they are brought into the lives of athletes.
The concept of ethical technology use that I discovered in my STS thesis could be applied to the projects that I completed in my internship with Capital One. In my Technical Report, I discuss the three projects that I completed throughout the summer. The first project was to complete memory and runtime testing for the company-created machine learning metrics that they used to analyze results of models. In completing this project, I was able to report to my team any bugs in the code as well as help them update their documentation with the results of the testing so that developers could decide which metrics were best to use for various dataset sizes or project types. The second project that I completed was a use case project. I was to take the same set of metrics and answer a set of questions so that they could update the metrics and documentation in the next version of them if I came across any bugs, difficulties, or found ways to improve the code. The final project was a User Interface project to create a portfolio of all of my involvement with Capital One over the course of my internship and display it in a similar format to how they display the metrics to those who do not code. The purpose of this was to understand how they use front end tools to help business analysts use their code without having the knowledge of coding that a software engineer would have. Overall, these tasks created an incredible learning experience for me as well as showed me that with my experience in classes from UVA, I would be able to join a team in the real world after college with confidence just as I had done in my internship.
BS (Bachelor of Science)
athletics, sports, machine learning, artificial intelligence
School of Engineering and Applied Science
Bachelor of Science in Computer Science
Technical Advisor: Rosanne Vrugtman
STS Advisor: Hannah Rogers
Technical Team Members: N/A
All rights reserved (no additional license for public reuse)