Abstract: Data privacy and regulations prevent the free transmission and sharing of information to a central place. While these regulations aim to ensure data owners can maintain clear control of their data, they also inhibit the training of machine learning (ML) algorithms and the analysis of processes that would benefit multiple stakeholders. Federated learning (FL) has provided an alternative to ensure multiple data owners collaboratively train ML models without sharing their data with each other. Sharing knowledge, rather than data, has allowed the mitigation of some of data privacy exposure risks. However, there are multiple open challenges that need to be addressed. In this talk, I will provide an overview of the state of the art and open challenges. Some of the topics I will discuss include the impact of FL on inference of private data, fairness, and transparency.

Bio: Nathalie Baracaldo leads the AI Security and Privacy Solutions team and is a Research Staff Member at IBM’s Almaden Research Center in San Jose, CA. Nathalie currently focuses on two main areas: federated learning, where models are trained without directly accessing training data and adversarial machine learning, where defenses are designed to withstand potential attacks to the machine learning pipeline. Nathalie is a technical lead for IBM federated learning. She also received the 2021 Corporate Technical Recognition, one of the highest recognitions provided to IBMers for breakthrough technical achievements that have led to notable market and industry success for IBM. This recognition was awarded for Nathalie's contribution to the Trusted AI initiative. In 2020, Nathalie received the IBM Master Inventor distinction for her contributions to the IBM Intellectual Property and innovation. She has published more than twenty papers in peer-reviewed conferences and journals, receiving four best paper awards. Nathalie received her Ph.D. degree from the University of Pittsburgh, USA in 2016.