Multidisciplinary data analysis in Big-data: from High Energy Physics to Astrophysics and Medical Sciences

• Grant number: 23/13749-1
• Support Opportunities: Research Projects - Thematic Grants
• Start date: May 01, 2025
• End date: April 30, 2030
• Field of knowledge: Interdisciplinary Subjects
• Principal Investigator: Jun Takahashi
• Grantee: Jun Takahashi
• Host Institution: Instituto de Física Gleb Wataghin (IFGW). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• Pesquisadores principais:( Atuais ): Donato Giorgio Torrieri ; Matthew William Luzum
• Pesquisadores principais:( Anteriores ): Tobias Frederico
• Associated researchers: André Veiga Giannini ; Cristiane Jahnke ; David Dobrigkeit Chinellato ; Dean Lee ; Donato Giorgio Torrieri ; Fernando Gonçalves Gardim ; Gabriel Silveira Denicol ; Jean Francois Paquet ; Joao Batista Florindo ; Jorge José Leite Noronha Junior ; José Barreto Campello Carvalheira ; Luiz Vitor de Souza Filho ; Manuel Calderon de la Barca Sanchez ; Maria Carolina Santos Mendes ; Maria Emilia Seren Takahashi ; Michael Annan Lisa ; Tiago Jose Nunes da Silva ; Tobias Frederico

Abstract

The project aims to support the establishment and consolidation of interdisciplinary research groups dedicated to technological development and the analysis of Big Data, with application of state-of-the-art data analysis techniques, including Multi-Variation Methods (MVA) and Machine Learning (ML) to address complex challenges within four distinct areas: Experimental High-Energy Nuclear Physics (ALICE-LHC), Phenomenology (EXTREME), Astrophysics (VERITAS), and Medical Sciences (Biomarkers in Oncology).Advances in data science have the potential to catalyze and enhance scientific studies in several areas, but to have solid results and development of technological innovation, it is imperative to adhere rigorously to scientific methodologies and criterias. It is in this context that we propose the formation of multidisciplinary research collaborations to promote data analysis in several existing problems using and sharing new computational tools, in particular with the use of ML and MVA.The project initially includes four lines of research:In high energy nuclear physics, we propose to contribute to the analysis of data from the ALICE experiment at the LHC, using ML in the study of event reconstruction and heavy flavored particles in heavy ion collisions (PbPb). We will also check the possibility of using ML techniques in global analysis in search of rare and unique events. In the same area there will also be a focus on phenomenology, where we can apply analysis tools to theoretical simulations of collisions at existing colliders as well as a new accelerator under construction in the United States, the EIC (Electron Ion Collider). Our group (with support from the previous thematic project) has developed a sophisticated simulation chain that has been successfully used to study various effects measured at the LHC and RHIC. In this new project, we will continue these studies, designing new measurements and studying new physical properties. A particular focus will be (low-energy) nuclear structure and its effects on ultra-relativistic nuclear collisions, and the extraction of nuclear properties from collision data.The same analysis techniques developed for high energies can be adapted and applied in other fields, particularly in astrophysics. In this project, we intend to apply HBT correlation analysis, which is used in heavy ion collisions, to data from VERITAS observatory, to study astrophysical sources. The results of this analysis can also be applied in the future astrophysical observatory CTA (Cherenkov Telescope Array).Finally, in an existing collaboration, we have applied data analysis algorithms developed in physics to problems in medical sciences. We developed with ML a method to quantify the importance of clinical variables for a patient's outcome. With this project, we intend to expand these works with the use of data analysis methods with ML and MVA in new problems of oncology and nuclear medicine, including data from tomography images, radiomics and genomics.The general goal of the project is to ensure the effective training and qualification of researchers in data analysis across multidisciplinary domains. Our approach emphasizes the utilization of cutting-edge tools, including machine learning techniques, and the latest advancements in the field of data science. To achieve this, we place a strong emphasis on researcher development, offering specialized training programs and courses, facilitating knowledge sharing, disseminating results and algorithms.In scientific outreach, in addition to activities such as participating in extension events and giving lectures, organizing CERN Massterclass events in highschools, we are proposing to organize the translation and presentation of an IMAX scientific outreach film on high-energy collisions for elementary and high school students. In addition to dubbing it into Portuguese, we anticipate creating supplementary materials for middle and high school teachers. (AU)