Graph Neural Networks (click here to learn more)

A collaborative project between the University of St. Thomas - Houston and Rice University to establish a sustainable initiative to foster computational research, addressing the call to prepare a diverse an inclusive STEM workforce by increasing interest and competencies in undergraduate research.

This research involves the departments of Mathematics and Computer Science (UST), Biology (UST) and Computer Science (Rice) to address foundational questions in the area of large-scale graph processing using graph neural nets. Relational database technology will be adapted to perform distributed neural computations over huge graphs, which will be applied to tackle two problems in biology: 1) Metagenomic analyses of mosquitoes genetic sequences to discover viruses carried by mosquitos in Houston, and b) protein motif and structural analysis to discover cell functionality.

Metabolomics

This project addresses this question: can we adapt representations of computer code to characterize metabolic processes? How can we use those representations to better understand these processes? This idea stems from previous work in Big Code Analytics, under the hypothesis that methods developed for code repair and finding bugs and vulnerabilities in computer programs can be applied to metabolic processes.

Algorithmic Fairness

In collaboration with the Philosophy department at UST, in this project we are working on the creation of a framework for ensuring that Algorithms used in analytics, machine learning and Artificial Intelligence are created, trained, deployed and evaluated taking into consideration fairness and bias. This is a growing area of interest and our approach is Socio-technical in nature, that is, we consider all the contexts: technical, societal, ethical and legal.

C5-Celts Computing Club - City Challenges