← Search

MarcAurelio Ranzato

2 accepted papers

2022

Towards Learning Universal Hyperparameter Optimizers with Transformers

NeurIPS 2022accept

Meta-learning hyperparameter optimization (HPO) algorithms from prior experiments is a promising approach to improve optimization efficiency over objective functions from a similar distribution. However, existing methods are restricted to learning from experiments sharing the same set of hyperparame…

2021

Efficient Continual Learning with Modular Networks and Task-Driven Priors

ICLR 2021poster

Existing literature in Continual Learning (CL) has focused on overcoming catastrophic forgetting, the inability of the learner to recall how to perform tasks observed in the past. There are however other desirable properties of a CL system, such as the ability to transfer knowledge from previous ta…