Date of Award
5-2026
Culminating Project Type
Starred Paper
Styleguide
ieee
Degree Name
Computer Science: M.S.
Department
Computer Science and Information Technology
College
School of Science and Engineering
First Advisor
Adriano Cavalcanti
Second Advisor
Maninder Singh
Third Advisor
Khalil Jalal
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.
Keywords and Subject Headings
Benchmarking, Machine Learning, Gaussian, Random Forest, Deep Learning
Abstract
Electroencephalography (EEG) is widely used to classify cognitive and motor imagery objectives in brain-computer interface (BCI) systems. While machine learning frameworks such as TensorFlow, PyTorch, and JAX are common in deep learning research, the BCI community has not yet established a standardized benchmark for comparing classical and deep learning models for EEG classification tasks. This project proposes the first benchmark that compares Gaussian, RandomForest, and CNN-based deep learning models, where the models will be independently implemented in TensorFlow, PyTorch, and JAX using EEG datasets collected at the St. Cloud State University BCI laboratory. This study will evaluate accuracy, F1-score, training time, inference latency, GPU and memory usage, and implementation complexity. The results will provide practical recommendations for selecting machine learning frameworks for EEG/BCI applications. They will also introduce a reusable evaluation pipeline for future research.
Recommended Citation
Islam, Md Mahmudul, "Benchmarking Gaussian, RandomForest, and Deep Learning Models for EEG/BCI Classification Across TensorFlow, PyTorch, and JAX" (2026). Culminating Projects in Computer Science and Information Technology. 69.
https://repository.stcloudstate.edu/csit_etds/69
Comments/Acknowledgements
I would like to thank Dr. Adriano Cavalcanti for his time, guidance, and support as the chair of my starred paper committee. His patience and insights were very helpful throughout the development of this project.
I would also like to thank the other members of my committee, Dr. Maninder Singh and Dr. Jalal Khalil, for their time and for reviewing my work.
Finally, I would like to thank St. Cloud State University, and the faculty and staff in the Department of Computing, Informatics and Data Science, for providing the academic environment and resources that made this research possible.