Johanna Skåntorp

Doktorand på Matematiska Institutionen.

Forskningsområde: Tight, convex, and polyhedral embeddings of deep neural networks (DNNs).

Nyckelord: Optimization, mixed integer linear programing, copositive programming, semidefinite programming, polyhedral theory, DNN, adversarial training.

Handledare: Jan Kronqvist