Keynote Speaker

Prof. George Varghese (UCLA)

George Varghese is the Jonathan Postel Distinguished Professor in the Computer Science Department at UCLA. He is a member of the National Academy of Engineering and the Internet Hall of Fame, and a Fellow of the American Academy of Arts and Sciences and the ACM. He has won the Kobayashi and SIGCOMM Awards for his lifetime work in networking. In 2004, he co-founded a company called NetSift that was acquired by Cisco in 2005. His textbook Network Algorithmics was published by Morgan-Kaufman.

Keynote: AI + networking, and Where do Ideas come from?

A lifelong AI skeptic, I was captivated by GPT 3 and then bowled over by Claude 5.0. So in the last few years, I have wondered how AI – together with classical symbolic analysis – can help my field of network verification. I will describe some of our adventures at UCLA in this space starting with CoSynth that uses AI for configuration synthesis and a verifier in a loop. I will then describe 3 tools for testing protocol implementations like BIND and FRR. Eywa finds bugs in valid executions; CornerCase finds bugs caused by invalid inputs; finally SpecLens finds bugs caused by underspecification in RFCs. Each tool combines AI with classical methods (symbolic analysis for Eywa, constraint generation for CornerCase, and state machine modeling for SpecLens) to do better than the best AI models by themselves.

I will end by surveying three paradigms for generating ideas with examples related to networking.. First, I will talk about Polya’s Paradox of Generalization. Next, I will talk about Tiling Conceptual Spaces, and end with Confluences, a way to mix two intellectual streams. I will spend some time at the end talking about a new confluence I call AI + X, for all fields X, and state what I would like to see AI help with for networking and randomized algorithms.