Date: November 14, 2025
Title: Logic-Guided Machine Learning for Autonomous Network Management
Speaker: Maria Apostolaki, Princeton
Abstract: As computer networks grow faster, more complex, and increasingly critical to modern applications, managing them reliably has become a formidable challenge that exceeds the capabilities of manual, human-driven approaches. Machine Learning (ML) offers adaptability and scalability but lacks the guarantees, robustness, and interpretability required for high-stakes environments. Conversely, formal methods provide rigor and correctness guarantees but often falter in scalability and require exhaustive system models.
A promising direction is to blend ML with formal methods, combining the adaptability and scalability of ML with the rigor and guarantees of logic-based reasoning. This neurosymbolic paradigm holds the potential to transform network management systems into solutions that are not only automated, but also interpretable, verifiable, and robust. This talk will demonstrate the promise of the neurosymbolic approach through a few concrete examples and highlight the key challenges that make turning this vision into practice difficult. In particular, this talk will focus on two foundational roadblocks: how to extract symbolic representations of network knowledge, and how to integrate or enforce them within ML models.
Bio: Maria Apostolaki is an Assistant Professor of Electrical and Computer Engineering at Princeton University. Her research spans networking and security, with a focus on combining ML with formal methods for more trustworthy network management. She has received the NSF CAREER, Google Research Scholar Award, IETF/IRTF Applied Networking Research Prizes, and Commendations for Outstanding Teaching. Maria earned her PhD from ETH Zurich and was a postdoctoral researcher at Carnegie Mellon University before joining Princeton.