New Bridges between Mathematics and Data Science
Understanding the world through data and computation has always been at the heart of scientific discovery. In the past decade primarily data-driven approaches, such as neural networks, have been very successful. Nevertheless, the reason for this success is to some extent mysterious and raises multiple questions regarding the robustness, explainability, interpretability and fairness of the algorithms used. The answer to these questions is crucial when it comes to decision making.
In recent years, there has been increasing activity in building bridges between these new ideas and other well-established approaches based on models typically derived from first principles. However, the computational cost of the latter makes them unaffordable except in low dimensions, which is a limitation from which neural networks are exempt. Establishing solid connections between these two different points of view have already proved to be extremely fruitful.
The Spanish ‘Mathematical Strategic Network’ (‘Red Estratégica de Matemáticas’, REM) organizes a one-week workshop in 8-11 November 2021 in Valladolid, Spain, with the aim of bringing together researchers in mathematics, machine learning, and data science, to exchange ideas and progress in the construction of new bridges among these fields and make visible the work already done.
The REM is a network of mathematical research institutes which aims at fostering the international presence of the Spanish research in mathematics, the creation of synergies among the mathematical scientific community and the socioeconomic impact of Spanish mathematical research. It promotes the dissemination and transfer of mathematical technology, orienting research towards the needs of companies, industries and public administrations.