User behavior recognition method based on multi-task multi-view incremental learning

Abstract

The invention discloses a multi-task multi-view incremental learning-based user behavior recognition method, which comprises the following steps of S1, collecting activity posture data, dividing the data into a plurality of parts, determining the number of tasks and the number of views of each task according to personnel participating in training, and initializing parameters of each stage; s2, extracting features for the sensor data by using a multi-task multi-view deep neural network; s3, parameter updating: calculating a plurality of loss functions, and performing parameter updating through back propagation; and S4, finishing training in all stages, and obtaining a final prediction result. The method has the advantages that the MTMVIS uses each layer of the multi-task multi-view deep neural network to extract features, and uses the attention layer to weight the output of all layers of the multi-task multi-view deep neural network for each task to serve as a final output layer. The MTMVIS uses adaptive weight consolidation to mitigate catastrophic forgetting issues and enhance model scalability.

Description

technical field [0001] The invention belongs to the technical field of smart device user behavior prediction, and in particular relates to a user behavior recognition method based on multi-task and multi-view incremental learning. Background technique [0002] In recent years, user behavior recognition through intellisense has attracted more and more researchers' interest in both academic and industrial fields. Various sensors (such as accelerometers, gyroscopes, etc.) embedded in powerful smartphones or smart wearable devices are used to identify user behavior, and are widely used in medical services, commerce, security, and other fields. User behavior recognition aims to detect user behavior in the real world, which can allow intelligent systems to help individuals improve their quality of life in areas such as healthcare, smart cities, and more. In previous studies, relatively good results have been achieved for user behavior recognition on specific datasets. [0003] H...

AI Technical Summary

Problems solved by technology

[0003] However, the existing user behavior recognition methods have the following drawbacks: (1) The traditional method collects data from various sensors at all time intervals to build a general offline model to recognize activities, which consumes a lot of space to store large amount of training data

Traditional incremental learning methods often ignore these issues

Images