Visit hope.edu

Computer Science Seminar: Summer Research Student Presentations

Thursday, October 28 at 11:00am

VanderWerf Hall, 102
27 Graves Place, Holland, MI 49423-3617

“Applying Machine Learning to Tennis” by Kaley Wilson and John VerMeulen, mentor Dr. Ryan McFall

In this presentation, John and Kaley will describe their work utilizing machine learning techniques to automatically identify various actions in a video of a tennis match.  In particular, they will talk about applying convolutional neural networks (CNNs) to classify the type of shot contained within an individual video frame.  As part of their talk, they will describe using a framework named You Only Look Once (YOLO) to detect and label individual objects within a video frame, and using pre-trained CNNs to aid in the classification task.

“Applications of Machine Learning using SciKit-Learn and TensorFlow” by Kenneth Munyuza and Trevor Palmatier, mentor Dr. Fola Olagbemi

Machine Learning – Learning from Data – is a field of study that is being successfully applied in several disciplines and industries. The earlier part of the project focused on exploring some of the traditional regression and classification machine learning models and their applications, and implementing various approaches to improve the accuracy of prediction of the classification models or minimize the error (a measure of the model’s accuracy) of regression models. The latter part of the project, which was completed in collaboration with Dr. Brooke Odle (Engineering), entailed developing an artificial neural network (ANN) trained on data obtained from healthy subjects equipped with inertial Measuring Units (IMUs) while they stood on force plates and performed a series of tasks that simulate motions assumed by caregivers while performing patient-handling tasks. The ANN was used to estimate ground reaction forces typically measured by force plates to facilitate the analysis of mechanics related to performance of patient-handling tasks outside of a gait laboratory. 

“Lynnette Redesign: Data-Driven Redesign of an Intelligent Tutoring System (ITS) for Middle School Algebra” by Marcus Artigue, research was through Carnegie Mellon University in Pittsburg, Pennsylvania

Lynnette is an AI-based tutoring system for middle-school students learning equation solving. In past studies it has been found to be very effective in helping students learn. This work consists of a redesign of the tutor based on past user data with the goal of improving usability and optimizing student learning.

Mentors were Dr. Vincent Aleven and graduate student Tomohiro Nagashima from Carnegie Mellon Human–Computer Interaction Institute. I also worked with another undergraduate student, Michelle Ma from UCLA, on this project.

 

 

 

Event Type

Academics, Natural & Applied Sciences Division, Computer Science

Subscribe
Google Calendar iCal Outlook

Recent Activity