PhD Positions in Privacy-Preserving Distributed Artificial Intelligence

Two funded PhD positions are available in the area of “Privacy-preserving Distributed Machine Learning.” The PhD candidate will work under the supervision of Prof. Ferdinando Fioretto at the EECS Department, Syracuse University. The position start date is flexible, with a start date as early as January 2020.
The PhD candidate is committed to conduct independent and original research, to report on this research in international publications and conference presentations, and to describe the results of the research in a PhD dissertation.

Topic Description

The recent surge in optimization and machine learning research, in particular, deep learning, paved the way for a number of applications, many of which use privacy-sensitive user data. The resulting models have been shown to often reveal private user information, which may harm individual users.
To contrast these risks, a new line of research aims at developing variants of optimization and ML algorithms that preserve the privacy of the individuals contained in the used datasets. Additionally, there is an increasing interest in leveraging distributed data shared across organizations to augment AI-powered services. Examples include transportation services, sharing location-based data to improve on-demand capabilities, and hospitals, sharing data to prevent epidemic outbreaks. The proliferation of these applications leads to a transition from proprietary data acquisition and processing to data ecosystems where different agents learn and make decisions using data owned by different organizations, boosting the need for privacy-preserving technologies.
The project focuses broadly on protecting the privacy of individuals without losing the benefits of large scale data analysis. Topics of interest include:
  • Privacy-preserving technology, such as Differential Privacy and secure multi-party computation
  • Distributed Machine Learning
  • Privacy-preserving Multiagent Systems
  • Privacy-Preserving Adversarial Deep Generative Models
The project will combine fundamental aspects of privacy, optimization and distributed computation to design algorithms that perform (distributed) machine learning and decision making while guaranteeing they do not violate privacy.
The ideal candidate will have a strong background and interest in machine learning, privacy-preserving technologies, and/or multi-agent systems. Publications in leading international venues (such as AAAI, IJCAI, AAMAS, ICML, NeurIPS) will be an advantage.

To Apply

Applications should be submitted at and candidates should include their resume and transcript (if available).