Unsupervised Topological and Contrastive Representation Learning for Galaxy Morphological Analysis
Persistent-homology descriptors and contrastive/self-supervised representations for unsupervised analysis of galaxy morphology.
Talks & Posters
Selected presentations
Seminars, conference talks, and posters on topological data analysis, astrophysics, complex systems, and scientific machine learning. Videos and supporting material are linked when available.
Latest seminar · IAG-USP · September 3, 2026
Persistent Homology as a Multiscale Observable:
From Quasiperiodic Spectra to Stellar Variability
A multiscale view of persistent homology as a physical observable, moving from quasiperiodic spectra and microscopic structure to trajectories, time series, and stellar variability.
Conference talks
Applying Statistical Methods in Astrophysics: Analysis of the Period-Luminosity Relation for Cepheids
Statistical and Bayesian analysis of the Cepheid period-luminosity relation using Markov Chain Monte Carlo methods.
Posters
Topological Machine Learning for Predicting Spatiotemporal Evolution in 1D Magnetohydrodynamics
Topological diagnostics of MHD simulations designed to characterize failure modes of neural temporal predictors.
Topological Data Analysis for Gravitational Wave Detection under Low Signal-to-Noise Ratios
Persistent-homology summaries for simulated gravitational-wave-like signals in low signal-to-noise regimes.
Recorded seminars
Selected longer-form seminars with publicly available recordings.
Computational Topology and Complex Systems
An introduction to computational topology and persistent homology, followed by applications to nonlinear and complex systems.
Topological Data Analysis Applied to Complex Systems
An overview of topological data analysis and its use for characterizing complex physical systems across scales.