Canonical Network

Canonical networks represent a fundamental approach to simplifying complex systems by decomposing them into smaller, reusable building blocks. Current research focuses on developing algorithms, such as non-negative matrix factorization and self-supervised learning methods, to identify these canonical structures within diverse data types, including brain networks, 3D object representations, and graph structures. This work aims to improve interpretability, scalability, and predictive power in various fields, ranging from neuroscience and computer vision to network analysis and machine learning. The resulting insights offer potential for advancements in understanding complex systems and developing more efficient algorithms for data analysis and prediction.