The proposed program to study the mass transfer between subhalos and their host clusters, as well as constrain the cross-section of dark matter, has not yet been completed. However, significant progress has been made in several areas and innovative outcomes have been created from the work.
In preparation to ingest a large number of observations and their modelisations, I developed several analyses necessary to better standardise and analyse collected data. It resulted in the supervision of four students developing the code to 1) select galaxies with their red sequence, 2) identify from their luminosity distribution the coarse distribution of underlying dark matter distribution, 3) analysis of lensed galaxies' morphological properties to automatically find them in the data, and 4) increase the database of lens models. This work is the basis of the upscaling needed by the community to prepare for the transformative sky survey happening with the launch of the Euclid telescope this summer.
I further developed a technique to automate the modelling with a machine learning approach in collaboration with a postdoctoral researcher in computer science to develop new algorithms. However, my collaborators left academia to join the industry and this stopped the project mid-course and as of now, it has not been able to continue. However, all the developed codes and procedures have been shared and the database of training models is currently under development. Before it can be used with the upcoming sky surveys, further developments are still needed
I cosupervised Sergio Garcia, a PhD student at the University of Michigan and we are working to use machine learning techniques to detect the signatures of alternative dark matter models in a suite of cosmological simulations self-interacting dark matter. Recently, the project has made significant progress in publishing one of the first analyses using the JWST images and showing how much the intra-cluster light can trace dark matter and hint at their link to the structure formation history. Additionally, My leadership in large international collaborations, such as BUFFALO and the EUCLID consortium supported the publication of several publications on developing new techniques. Moreover, PyLenstool, a python wrapper for lens models, has been developed to increase batch modelling. Additionally, the project has developed a large suite of functions for simulations of models, in order to increase the training necessary to train machine learning techniques aiming to automate the modelisation.
Thanks to the development of new models and a deeper understanding of the current state of modelling, we built new selection methods for clusters following the growth of the structure. This evolutionary track has been at the basis of several observing proposals and culminated at the last call for space telescope time request of JWST by the holder of this fellowship and leading a collaboration of 23 research spread across nearly tens institutes in Europe and the USA.
Overall, the vision of the research developed in this proposed program is still ongoing but has delivered significant progress in terms of data collection and modelling. Further work is needed to fully achieve the objectives of deepening our understanding of the growth of galaxy clusters.
My work has been advertised to a large audience of experts in the field with my participation in the specific session of the European Week of Astronomy. I was invited last year at EAS 2022 and again this year at EAS2023. This is a direct result of the publication of my results through peer-reviewed journal/
In addition, I disseminate my results into dedicated meetings and collaboration meetings over the years with about 15 of those meetings in total.
During a collaboration meeting at International Space Science Institue (Bern) I disseminate the library of models that will be employed and made connections needed to support further it dissemination reaching several other institutes.
