CompactTag: Minimizing Computation Overheads in Actively-Secure MPC for Deep Neural Networks

In this talk, I presented my preliminary results on CompactTag, which significantly reduces the tag computation costs for ML with dishonest majority.