A Count-sketch to Reduce Memory Consumption when Training a Model with Gradient Descent.

Wissam Siblini, Frank Meyer, Pascale Kuntz

Browse the full IJCNN paper archive.