June 9, 10:00 AM
We are pleased to announce the upcoming PhD defense of Samie Mostafavi, who will present his thesis titled:
“Predictability, Prediction, and Control of Latency in 5G and Beyond: From Theoretical to Data-Driven Approaches”. In the lead-up to the defense, we asked Samie to share insights about his research, its societal impact, and the challenges that lie ahead.
What is your thesis about?
Wireless networks have become the ‘nervous system’ of a digital society. Everything from a video call on the subway to a fleet of autonomous drones depends on them. Yet their performance is governed by an intractable mix of fading radio channels, user mobility, and in-network queuing that makes packet delays feel random.
Samie’s research aims to bring order to this apparent randomness. His thesis introduces:
- Open tools and testbeds that bridge theory and practice, allowing researchers and operators to measure, analyze, and act on real 5G delay data.
- A theoretical framework to quantify how predictable wireless latency can be, and to identify the most relevant telemetry.
- Machine learning-based prediction models, from extreme-value techniques for rare delay spikes to Transformer models for real-time pattern tracking.
What impact does this work have on society?
When we can forecast wireless delay instead of guessing it, we unlock progress on three fronts:
- Everyday experience: Smoother video, quicker cloud gaming, more responsive XR and AR headsets, and fewer “buffer wheels” all come from networks that understand their own limits and allocate resources just in time.
- Critical services. In automated factories, data packets have to arrive within milliseconds. By understanding and predicting network delays, we can keep those packets on schedule and the machines running smoothly.
Are there any remaining challenges in this area?
“The key challenge now is scaling prediction from a single connection to an entire network.”
Samie highlights the difficulty of real-time prediction across thousands of 5G cells and stresses the importance of developing standard interfaces for large-scale integration — especially as we move toward 6G networks.
Mark your calendar and join us on June 9 at 10:00 AM to learn more about this exciting work and celebrate this significant milestone in Samie’s research journey.