Palestra: Machine learning for scientific discovery – How far can data-driven approaches take us?

  • Palestrante: Emille E. O. Ishida, pesquisadora do CNRS, França
  • Data: 29/05/2023
  • Horário: 13:00 às 14:00
  • Local: Sala S-006-0

The significant increase in volume and complexity of data resulting from technological development is a common challenge faced by many scientific disciplines. More sensitive detectors and large scale experiments are currently overwhelming researchers who were obliged to turn to automatic machine learning algorithms in order to filter, order or pre-select potentially interesting subsets of the original data for further scrutiny. In this scenario, algorithms need to be carefully designed to be flexible, selecting scientifically interesting/useful examples in accordance with the expert’s needs, thus optimizing the distribution of human efforts in scanning a large data set. In this talk I will show a few examples of machine learning applications to astronomy. I will describe the current state of the art and discuss the limits of data driven approaches in a scientific data environment, advocating for a personalization of machine learning algorithms dedicated to scientific exploration.

Biography: Emille E. O. Ishida is a research engineer at CNRS, based at Clermont Ferrand, France. She is co-founder of the Cosmostatistics Initiative (COIN) and the SNAD collaboration and is scientific PI of the Fink broker. She mainly works in machine learning applications to astronomy, with special emphasis on integration of expert knowledge in the learning cycle. She is also engaged in research for development of interdisciplinary scientific environments able to foster fruitful collaboration inspired by astronomy.

Link do evento