Feature extraction using multi-task learning

US11100399B2 · kind B2 · utility

Assignee

• International Business Machines Corporation · US

Inventors

• Wei Shan Dong · Beijing, CN
• Peng Gao · Nanhu, CN
• Chang Sheng Li · Beijing, CN
• Chun Yang Ma · Beijing, CN
• Kai AD Yang · Beijing, CN
• Ren Jie Yao · Beijing, CN
• Ting Yuan · Shanghai, CN
• Jun Zhu · Shanghai, CN

Key dates

• Filing date: Nov 21, 2017
• Grant date: Aug 24, 2021
• Expiry date: Jun 25, 2040

Classification

• Technology area (CPC G): Physics
• CPC primary: G06N3/048
• WIPO field: Computer technology
• WIPO sector: Electrical engineering

Abstract

Systems and methods for training a neural network model are disclosed. In the method, training data is obtained by a deep neural network (DNN) first, the deep neural network comprising at least one hidden layer. Then features of the training data are obtained from a specified hidden layer of the at least one hidden layer, the specified hidden layer being connected respectively to a supervised classification network for classification tasks and an autoencoder based reconstruction network for reconstruction tasks. And at last the DNN, the supervised classification network and the reconstruction network are trained as a whole based on the obtained features, the training being guided by the classification tasks and the reconstruction tasks.