Search Results for author:
Found 4 papers, 1 papers with code
CARLS: Cross-platform Asynchronous Representation Learning System
In this work, we propose CARLS, a novel framework for augmenting the capacity of existing deep learning frameworks by enabling multiple components -- model trainers, knowledge makers and knowledge banks -- to concertedly work together in an asynchronous fashion across hardware platforms.
A Simple and Effective Positional Encoding for Transformers
Our analysis shows that the gain actually comes from moving positional information to attention layer from the input.
Adversarial Robustness Across Representation Spaces
In this work we extend the above setting to consider the problem of training of deep neural networks that can be made simultaneously robust to perturbations applied in multiple natural representation spaces.
Finding Fast Transformers: One-Shot Neural Architecture Search by Component Composition
Transformer-based models have achieved stateof-the-art results in many tasks in natural language processing.
