Electronic component library management method based on ai

Abstract

The application relates to the technical field of artificial intelligence, and provides an AI-based electronic component library management method, which can perform continued mask language model training on a BERT layer in BERT-BiLSTM-CRF, can enhance field data adaptability, can reduce the difficulty of subsequent training, can improve context semantic understanding capability, can perform multi-granularity feature fusion based on a multi-granularity prior knowledge fusion attention mechanism, can make the model more comprehensive and focus on key features, and can improve model accuracy; can perform entity extraction on a target material description text by using an electronic component entity extraction model, can break the dependence on fixed delimiters and hard coding rules, can process material description texts in any format, can solve the problem of component description fragmentation and term-intensive adaptability; and can update a library field incrementally to an electronic component library, so that automatic and efficient management of the electronic component library is realized.

Images