Sentence pair intelligent semantic matching method and device for judicial public service

Abstract

The invention discloses a sentence pair intelligent semantic matching method and device for judicial public service, and belongs to the technical field of artificial intelligence and natural languageprocessing. The technical problem to be solved by the invention is how to capture more semantic context features, the relationship of coded information between different dimensions and the interactioninformation between sentences, and intelligent semantic matching of sentence pairs for intelligent judicial public service is realized. The adopted technical scheme is as follows: a sentence pair semantic matching model consisting of a multi-granularity embedding module, a time sequence feature convolutional network construction module, a feature conversion network module and a label prediction module is constructed and trained; and time sequence feature convolution representation of sentence information and two-dimensional convolution coding representation of semantic features are achieved,meanwhile, a final matching tensor of sentence pairs is generated through an attention mechanism, and the matching degree of the sentence pairs is judged, so that the purpose of intelligent semantic matching of the sentence pairs is achieved. The device comprises a sentence pair semantic matching knowledge base construction unit, a training data set generation unit, a sentence pair semantic matching model construction unit and a sentence pair semantic matching model training unit.

Description

technical field [0001] The invention relates to the technical fields of artificial intelligence and natural language processing, in particular to a sentence-pair intelligent semantic matching method for judicial disclosure services. Background technique [0002] In recent years, the Supreme People's Court has successively promulgated a series of institutional regulations on judicial openness, continuously strengthening the implementation of open case filing, open court hearings, open trial results, open judgment documents, and open execution processes, so as to promote justice through openness. In order to further promote judicial openness and satisfy the people's right to know and right to participate, it is very important to promptly respond to the public's inquiries about judicial openness services. Facing the ever-increasing number of judicial consultation requests, how to respond quickly and accurately is an urgent problem that judicial institutions at all levels need t...

AI Technical Summary

Problems solved by technology

The cyclic neural network adopts a chain structure. Although it can capture long-distance semantic features well, it does not fully capture local features.

This will lead to the loss of some local semantic information in the sentence, making the captured semantic features incomplete

In addition, due to the impact of its chain structure, that is, the state of the next time step depends on the operation result of the previous time step, which leads to its low execution efficiency

In contrast, the convolutional neural network can effectively capture local information and has good parallelism; however, due to the limited size of the convolution kernel of the convolutional neural network, the temporal features it can capture are relatively limited. limit

Both the recurrent neural network and the traditional convolutional neural network have some disadvantages that cannot be ignored in the encoding and processing of sentence semantic information.

Images