Welcome
I am Stratos, a Visiting Scholar at the CGU Institute of Mathematical Sciences, where I collaborate with Distinguished Research Professor Hrushikesh Mhaskar on computational harmonic analysis and machine learning, supported by the US Office of Naval Research.
I completed my PhD in Mathematics at the University of Maryland in 2024 under the supervision of Radu Balan. My dissertation, G-Invariant Representations using Coorbits, develops machine learning models for graphs that are invariant to node relabelings. Before that, I received my BS (2016) and MS (2019) in Mathematics from the University of Crete, advised by Themistoklis Mitsis.
Research interests
My research lies in applied harmonic analysis and geometric machine learning. Specific topics include:
- Group-invariant representations — embeddings of orbit spaces under group actions (permutations, $O(d)$), injectivity and bi-Lipschitz properties.
- Phase retrieval — connections between phase retrieval and permutation-invariant embeddings.
- Kernel methods and approximation theory — local kernel expansions, classification from a signal-separation viewpoint, function approximation by neural networks.