The team was tasked with creating a program that can compress data dynamically given the bandwidth and CPU usage so that neither are at a bottleneck and the optimal amount of data is being sent. Research area: Systems.
Clients waste time finding qualified and available contractors in their area, while contractors have difficulty getting leads that materialize into clients. Altidor is a web-based platform that match clients to qualified contractors, including search and scheduling algorithms, and a payment portal. Research area: AI, ML, NLP.
Using user flight history data from American Airlines, the team created a machine learning model that recommends a user’s next destination. The model performs item-based collaborative filtering through a K-Nearest-Neighbors model to find the most similar destinations based on travel history. Research area: AI, ML, NLP.
The team created an algorithm that efficiently searched through 4K to 12K resolution images in a store’s security footage to improve speed (24 frames per second) and accuracy (retained details of people at a distance). The algorithm used image slicing, model optimization and compression. Research area: Computer vision.
The Bray International Health Analytics platform is designed to provide valve health information for use in planning pro-active maintenance and repairs, saving users millions of dollars in savings by reducing component downtime. Research area: Data science.
Our project uses curriculum learning to train a machine learning agent to play a game of “capture the flag” in a two vs. two game simulation. The team surveyed different “temporal and reward phasing” techniques using different centralized and decentralized variants. Research area: AI, ML, NLP.
The goal of the team's project was to create a glTF geometry importer and to improve upon the ray tracer that connects BRL-CAD and the production rendering of appleseed. Research area: AI, ML, NLP
This project aims to improve the user experience and functionality of the physically based rendering component of the engineering CAD system BRL-CAD. Research area: Graphics, visualization and computational fabrication.
A team of six students optimized an algorithm and augmented a dataset to improve a machine learning algorithm for detecting breakdowns in people with autism. Research area: AI, ML, NLP.
Online marketplaces are difficult to navigate due to products classified in the wrong category. The team created an algorithm which identified products incorrectly classified, based on the absence of unique phrases which characterize a category of products, in a product’s description. Research area: AI, ML, NLP.
A team of five designed and implemented a system that can rate the accuracy of a product’s placement in a given classification taxonomy. The project used machine learning and natural language processing techniques. Research area: AI, ML, NLP.
A team of five designed a system that measures both the correctness and precision with which products are classified in a hierarchical taxonomy using outlier detection. Research area: AI, ML, NLP.
The project created a machine learning model to assess product categorization to uncover errors in product taxonomy labels. The team used different inputs like product image descriptions to calculate a confidence score, indicating the correct classification of a product within each level of the taxonomy. Research area: AI, ML, NLP.
The team created an algorithm that efficiently searched through 4K to 12K resolution images in a store’s security footage to improve speed (24 frames per second) and accuracy (retained details of people at a distance). The algorithm used image slicing, model optimization and compression. Research area: Computer vision.
Entire teams of engineers repeatedly develop model-view-controller applications for each specific use case, from aircraft systems to desktop computers. VizUI automates this time-consuming process by automatically generating a model-view-controller web application from a Qt Creator UI file instantly. Research area: Human-computer interaction.
