Dr. Siddiqua Mazhar

Degrees and Credentials

Ph.D. Mathematics, Newcastle University, UK 2017

Short Bio

Siddiqua Mazhar holds a Ph.D. in mathematics in the field of Computational Group Theory from Newcastle University, UK. Her academic journey led her to transition into the field of artificial intelligence and neural networks, where she honed her skills as a Data Engineer at the University of Arizona for a year. Her passion for teaching led her to serve as an Adjunct Professor at both Mid Michigan Community College and Southern New Hampshire University.

Currently, Siddiqua is deeply involved in multiple projects, with a particular focus on Deep learning from a signal processing perspective. These projects aim to unravel the extraordinary success of deep learning in real-world applications, where cutting-edge machine learning techniques rely heavily on neural networks. Despite their remarkable performance, the precise mathematical theory explaining their effectiveness remains elusive.
In these endeavors, Siddiqua introduces a groundbreaking mathematical framework that delves into the intricate functional properties of neural networks trained to fit data. The framework draws on essential mathematical tools, including transform-domain sparse regularization, the Radon transform of computed tomography, and approximation theory, which all have their roots deeply embedded in signal processing.

Within this innovative framework, Siddiqua sheds light on critical aspects of neural network training, uncovering the impact of weight decay regularization, the significance of skip connections and low-rank weight matrices in network architectures, the role of sparsity, and the underlying reasons behind the exceptional performance of neural networks in handling high-dimensional problems. Her work represents a significant step towards a more comprehensive understanding of deep learning and its remarkable achievements in the realm of artificial intelligence.

Besides with her expertise in ELT (Extract, Load, and Transform) processes, Python, Shell scripting, SQL (MySQL, AWS Redshift), NoSQL (MongoDb), Power BI, Docker, Jira project management, Git version control, AWS S3 data storage, and Machine Learning, she is equipped to excel in various computational domains. Her diverse skill set allows her to streamline data workflows, develop scalable solutions, create insightful visualizations, manage projects efficiently, ensure code integrity, and apply advanced data analysis techniques. Continuously exploring emerging technologies, she is a valuable asset in driving innovation and delivering impactful results across diverse projects and industries.

Academic Focus

Dr. Mazhar taught Calculus I, Calculus II, Calculus III, College Algebra, Linear Algebra, Abstract Algebra, Statistics and Probability with Python, Introduction to Topology, Differential Equations, Group Theory, and Semigroup Theory.

Research, Accomplishments, and Publications

Mazhar, S. (2020). Composition of permutation representations of triangle groups. Communications in Algebra. 2020, 48(2), 792-802 (DOI: 10.1080/00927872.2019.1662911)