In order to realize efficient and accurate intellectual property infringement detection,this paper studies the comprehensive analysis method of multi-mode heterogeneous information driven by big data.To obtain massive multi-source data by means of distributed crawler and API collection,NLP,image and CNN to construct the relationship between data.Finally,cloud native deep learning model is built to realize end-to-end training and integration of multi-modal features.The system supports the collection,representation,modeling and analysis of heterogeneous data,and can be easily integrated into mobile terminals and Web applications.The results show that the accuracy of the method is more than 90%,which is higher than the single data source and model method.This study provides a useful reference for constructing an efficient IP protection system.
intellectual propertyinfringement detectionmultimodal datadeep learningbig data
维普
万方数据
