Towards modular and programmable architecture search
We propose a formal language for encoding search spaces over general computational graphs, applicable in particular to neural network architecture search.
Extreme Dimensionality Reduction for Network Attack Visualization with Autoencoders
We used semi-supervised Autoencoders to obtain 2d visualizations of network traffic that separate between distinct types of attacks.
A meta-analysis approach for feature selection in network traffic research
We analyse the used features in network traffic research, and propose a new traffic vector based on how often they are chosen in the literature.
Jointly Learning to Embed and Predict with Multiple Languages
We propose a joint formulation for learning task-specific cross-lingual word embeddings, along with classifiers for that task. We obtain state of the art results in multiple multilingual datasets.