A Broad Learning System Based on the Idea of Vertical Federated Learning

AbstractAs a new type of deep neural network, the broad learning system has been attracting attention since it was proposed. It shows good performance in image processing, face recognition, etc. To train a broad learning model requires comprehensive data. But the data on which the models are trained is often held by different companies and these companies are reluctant to share the data. To address this problem, we propose a broad learning system based on Vertical Federated Learning. Compared with the existing broad learning systems, it can integrate multi-party data to obtain comprehensive information, and realize the training of broad learning systems in the case of isolated data islands. This facilitates the application and advancement of broad learning systems.KeywordsBroad learning systemVertical federated learningDeep neural network