Vector Institute Shares ICML Research on Multi-Agent and Privacy AI

VectorInst · x · 2026-07-10

Vector Institute shared various research presented by its members at the ICML conference, covering AI attribution methods via efficient Shapley value approximation, trust and cooperation mechanisms in multi-agent systems, as well as privacy and security directions like verifiable differential privacy and automated multi-agent red teaming.

Related event: Vector Institute Showcases Multi-Agent and Privacy AI Research at ICML(2 posts)→

Original post →

More from Research

Research channel →