Mathematical and Machine Learning-based Modeling for Brain Signal Analysis

Summary

This project is the continuation of the XSEDE Empower project "Computing and Mathematical Foundations for Brain Signal Analysis", which is being carried out during Summer 2021. The project aims to develop mathematical and machine learning-based models for brain signal analysis. Its focus is to analyze EEG brain signals for patterns that represent intended communication of non-verbal and/or paralyzed subjects. This project will include training on Machine Learning techniques at intermediate and advanced level for undergraduate students, who participated on the previous project.

Job Description

The student will perform the following tasks: 1) Learn and employ intermediate and advanced Machine Learning techniques for EEG brain signals analysis such as Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Autoencoders (AE), etc. 2) Develop mathematical and Machine Learning models for predicting intended subject communication by analyzing EEG brain signals. 3) Write and present reports, papers and/or presentations on the accuracy of the models.

Computational Resources

Students will use XSEDE resources for training, validating and testing the Machine Learning models. Expanse at SDSC and Bridges 2 at PSC are the two resources that will be used.

Contribution to Community

After participating in this project undergraduate students will be able to use XSEDE resources productively and to perform research with minimal supervision. In addition, as students continue their academic or professional endeavors, they will advocate for supporting XSEDE and its community.

Position Type

Apprentice

Training Plan

The project includes training on intermediate to advanced Machine Learning methods and applying them to develop models. In addition, students will learn/review how to perform research (literature review, develop an hypothesis, describe a methodology, carry out experiments, analyze results and provide conclusions). Weeks 1-4: 1) Learn Machine Learning methods, and develop models for identifying patterns in EEG brain signals Weeks 5-8: 2) Students will perform experiments using XSEDE resources and evaluate the developed models. Weeks 8-12: 3) Prepare report, paper and/or presentation about the findings on the accuracy of the developed models.

Student Prerequisites/Conditions/Qualifications

The minimum requirements are: The student must have participated in the XSEDE Empower project "Computing and Mathematical Foundations for Brain Signal Analysis" and/or have an equivalent background in mathematics, computing and EEG brain signals analysis.