Ramtin Pedarsani

Ramtin Pedarsani is an Associate Professor in the Department of Electrical and Computer Engineering at the University of California, Santa Barbara, where he leads a research group focused on machine learning and networked systems. He joined UCSB in Fall 2016 following postdoctoral work at UC Berkeley.

Education:

• PhD in EECS, University of California, Berkeley (2015)
• MSc, École Polytechnique Fédérale de Lausanne (2011)
• BSc, University of Tehran (2009)

His research spans machine learning, optimization, information theory, and intelligent transportation. Current projects investigate federated learning under communication bottlenecks, adversarial robustness against sparse attacks, fairness in sequential decision-making, reinforcement learning from imperfect demonstrations, and mixed-autonomy traffic control. His work emphasizes fundamental limits and efficient algorithms for real-world systems.

Recent publications (2022-2025) demonstrate strong focus on robustness in distributed ML (federated learning security, adversarial defenses) and intelligent transportation (multi-agent coordination, congestion-aware systems). Theoretical contributions in sparse attacks and fairness are frequently validated through applications in autonomous networks.

Honors:

• IEEE ICC Best Paper Award (2014)
• Joint ComSoc/ITSoc Best Paper Award (2020)
• NSF CRII Award (2017)

He advises 6 PhD students working on reinforcement learning, robust ML, and fairness. Former PhD advisees hold positions at Meta and MIT. His group collaborates with industry partners on federated learning and autonomous systems, with grants including NSF support for privacy-preserving distributed computation.

Leads the UCSB Learning and Adaptive Systems Lab, investigating fundamental ML principles applied to networked infrastructure. The team develops algorithms for federated optimization, adversarial robustness, and transportation networks with mixed autonomy.