work experience and research

I worked on a new approach for galaxy morphological classification, consisting in feature-extraction from 2D-images via isophotal analysis of the galaxy-light distribution. Features are organized as 1D sequences including enough information to successfully distinguish between early-type and late-type galaxies. The sequences are processed through Modulos AutoML to perform automated model selection and training. Dimensionality reduction implies faster data manipulation and less computational cost, without significant loss in accuracy. This method can be potentially used to detect contaminating sources and for objects deblending, as well as for astronomical spectroscopy. Publications: Tarsitano et al., 2021, Image feature extraction and galaxy classification: a novel and efficient approach with automated machine learning (featured also on Modulos AI).

​

Leading the development of a novel project called PyCosmo, an interactive Python-based framework providing analytical and numerical predictions for cosmological quantities. It is publicly available online on the platform PyCosmoHub, https://pycosmohub.com/hub/login, where the users can work on Jupyter notebooks and benefit from easy tutorials and documentation. PyCosmoHub is useful both for educational and research purposes. Publications: Tarsitano et al, 2020, Predicting Cosmological Observables with PyCosmo, submitted to Astronomy and Computing. Additional developments currently focus on the implementation of new theoretical models to compute cosmological observables (e.g. the Kinetic Field Theory (KFT), described in Bartelmann et al., 2019).