Academic literature opinion evidence retrieval system and method based on local large model

Abstract

The application discloses a kind of academic literature viewpoint evidence retrieval system and method based on local large model, belong to natural language processing and academic literature retrieval technical field.The system of the present application includes front-end interaction module, SpringBoot back-end service module, PDF literature batch analysis module, local large model inference module, credible constraint control module, academic literature batch import module, viewpoint semantic retrieval module and source tracing and citation annotation module;The system uses local private deployment architecture, blocks data interaction with external public network;Credible constraint control module fixes local large model temperature parameter to 0.0, and generates answer based on local literature library content by closed prompt word limit model;Viewpoint semantic retrieval module carries pre-training semantic embedding model, and retrieves matching literature evidence in local academic literature library by vector similarity algorithm.The present application realizes academic literature semantic retrieval, zero illusion reasoning and standard citation generation in local environment.