Boris Knyazev

Graph Neural Networks for Learning Equivariant Representations of Neural Networks

Can We Learn Communication-Efficient Optimizers?

Learning Optimizers for Local SGD

Pretrained language models to solve graph tasks in natural language

Learning to Optimize with Recurrent Hierarchical Transformers

Can We Scale Transformers to Predict Parameters of Diverse ImageNet Models?

Hyper-Representations as Generative Models: Sampling Unseen Neural Network Weights

Pretraining a Neural Network before Knowing Its Architecture