Mathematical Analysis of Adversarial Machine Learning (25w5469)

The Casa Matemática Oaxaca (CMO) will host the "Mathematical Analysis of Adversarial Machine Learning" workshop in Oaxaca, from August 17 to August 22, 2025.

In the rapidly evolving landscape of machine learning and artificial intelligence, the pervasive integration of learning models into domains such as autonomous driving, medical diagnosis, drug development, and generative systems has highlighted a critical concern: the vulnerability of these models to adversarial attacks. These are imperceptible perturbations to algorithms' inputs intended to fool an algorithm to produce undesirable outputs. Most notably, as observed in the last years by researchers, state of the art learning models used for image classification are especially prone to adversarial attacks. This has hindered the safe application of such algorithms for the previously mentioned applied domains.

In this 5-day workshop, researchers from diverse fields such as mathematics, statistics, computer science, and data science will come together to thoroughly investigate the intricate connections within the realm of adversarial machine learning. The primary goal of the workshop is to enhance the theoretical foundations of this field, shedding light on existing defense mechanisms, and exploring new directions for proposing theoretically grounded defense strategies.

The Casa Matemática Oaxaca (CMO) in Mexico, and the Banff International Research Station for Mathematical Innovation and Discovery (BIRS) in Banff, are collaborative Canada-US-Mexico ventures that provide an environment for creative interaction as well as the exchange of ideas, knowledge, and methods within the Mathematical Sciences, with related disciplines and with industry. The research station in Banff is supported by Canada's Natural Science and Engineering Research Council (NSERC), the U.S. National Science Foundation (NSF) and Alberta's Advanced Education and Technology.