Distributed deep learning is an important and indispensable direction in the field of research. Earlier research has proposed many algorithms or techniques on accelerating distributed neural network training. This study discusses a new training scenario, namely data isolated learning. Specifically, each node its own local cannot be shared for some reasons. However, order to ensure generalizatio...

We present in this paper the main applications of data mining techniques at Electricité de France, the French national electric power company. This includes electric load curve analysis and prediction of customer characteristics. Closely related with data mining techniques are data warehouse management problems: we show that statistical methods can be used to help to manage data consistency and...

Because the majority of patients with chronic hepatitis C do not respond to interferon, alternative treatments need to be established. Several lines of evidence suggest that iron depletion is beneficial for such patients. Thus, gastrectomized patients with a reduced capacity for iron absorption might have an advantage in treatment of their liver damage over patients with intact gastrointestinal...