I am a machine learning researcher with a Ph.D. in Mathematics specializing in Graph Theory, working at the intersection of mathematics, machine learning, and interpretable AI.
My research focuses on understanding and improving the foundations of learning algorithms—not only implementing models. I have worked on neural network optimization and data efficiency, investigating how learning systems can achieve strong performance while using computational and training data more effectively. I am currently focused on explainable and interpretable machine learning.
With modern AI increasingly capable of generating code, I believe the ability to understand the mathematics, algorithms, assumptions, and optimization principles behind a model is becoming even more important. My goal is to build AI systems that are not only effective, but also efficient, interpretable, and technically well understood.
| Area | Expertise |
|---|---|
| Mathematical Foundations | Graph Theory, Mathematical Modeling, Algorithmic Reasoning |
| Machine Learning | Regression, Classification, Clustering, Feature Engineering, Model Evaluation |
| Deep Learning | Neural Networks, CNNs, Graph Neural Networks, Representation Learning |
| Neural Network Efficiency | Optimization, Data Efficiency, Training Efficiency |
| Explainable AI | Prototype Learning, Interpretable Decision Trees, Concept-Based Models |
| Computer Vision | Visual Regression, Feature Analysis, Interpretable Visual Representations |
| Graph Machine Learning | Graph Neural Networks, Graph-Based Learning, Molecular Property Prediction |
| Scientific Machine Learning | Thermal Data Analysis, Spatiotemporal Feature Engineering |
| Programming & Data | Python, PyTorch, NumPy, Pandas, scikit-learn |
Neural network optimization and data-efficient learning
More training data is not always better: identifying and removing less informative samples can improve accuracy, generalization, and robustness while increasing data efficiency, including in continual learning settings.
Focus: Data Efficiency · Sample Selection · Continual Learning · Robustness
Machine learning on 3D-printer thermal data to automatically detect abnormal thermal behavior and support preventive prediction of process anomalies in additive manufacturing.
Focus: Predictive Maintenance · Anomaly Detection · Thermal Data · Machine Learning · Additive Manufacturing
- Mathematical Foundations of Machine Learning
- Explainable and Interpretable AI
- Neural Network Optimization
- Data-Efficient Learning
- Prototype-Based Learning
- Graph Neural Networks
- Computer Vision
- Regression
- Scientific Machine Learning
My background in Graph Theory shapes how I approach machine learning problems: by looking beyond model APIs and implementation details to understand the underlying structures, optimization objectives, and algorithms.
My work spans three complementary directions:
Efficiency — How can neural networks learn effectively with fewer resources or less data?
Explainability — How can we understand why a neural network makes a particular prediction?
Mathematical Structure — How can mathematical reasoning help us design, analyze, and improve learning algorithms?
I am interested in research and engineering problems where these areas intersect—building machine learning systems that are mathematically grounded, computationally efficient, and interpretable.
- Strong mathematical and algorithmic problem-solving background
- Experience designing and evaluating machine learning research
- Neural network optimization and data-efficiency research
- Explainable and interpretable AI
- Graph-based machine learning
- Deep learning and computer vision
- Scientific data processing and feature engineering
- Experimental design and model evaluation
- Ability to translate mathematical and research ideas into working implementations
My research spans mathematics, machine learning, neural network efficiency, explainable AI, graph learning, and scientific applications of machine learning.
Google Scholar · Personal Website
Portfolio: https://MahsaMozafariNia.github.io LinkedIn: https://www.linkedin.com/in/mahsa-mozafarinia-66b5151a3/ Google Scholar: https://scholar.google.com/citations?user=jJUYsAgAAAAJ&hl=en