Event evolution reasoning prediction method based on cognitive driven intelligent generation

By constructing an interactive thought chain and a causal reasoning network, and combining cognitive logic knowledge with a false alarm detection model, the accuracy and reliability issues of traditional event evolution prediction methods are solved, achieving highly reliable event evolution prediction.

CN122334500APending Publication Date: 2026-07-0310TH RES INST OF CETC
0 Cites 0 Cited by

Patent Information

Application Number
CN202610493984.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-04-15
Publication Date
2026-07-03

AI Technical Summary

Technical Problem

Traditional event evolution prediction methods are insufficient in accuracy and adaptability when dealing with unforeseen or sudden events, making it difficult to provide forward-looking guidance. Furthermore, prediction methods based on open heuristics have reliability issues, and the generated prediction conclusions are prone to being inconsistent with the facts.

Method used

By leveraging a domain-wide language model to extract expert knowledge points, analysis processes, and logical reasoning, an interactive thought chain and causal reasoning network are constructed. This is combined with cognitive logic knowledge and factual prompts for the event to be predicted to perform event evolution reasoning and prediction. Finally, credible conclusions are screened through closed-loop evidence chain analysis and a false detection model.

Benefits of technology

It improves the accuracy and credibility of event evolution prediction by guiding a large language model to generate logically cognizable prediction conclusions and enhancing the robustness and reliability of the prediction conclusions through multi-level verification and screening.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses an event evolution reasoning prediction method based on cognitive-driven intelligent generation, and the method comprises the steps: firstly extracting expert knowledge points, an analysis process, an analysis view angle and affair logic from an event research report through a domain large language model, constructing an interactive thinking chain, extracting causal affairs, and constructing a causal affair network structure; the cognitive logic knowledge and the problem-related facts of the to-be-predicted event are fused and embedded into a prompt template to construct a prompt model, so that the field big language model generates an event evolution reasoning prediction result according to expert analysis dimensions and analysis processes, and finally conflict detection is performed on event elements through evidence chain closed-loop analysis. And performing credible conclusion screening to obtain a credible prediction conclusion. A prompt template for guiding a large model is constructed through cognitive logic learning, and multi-level credibility verification screening is combined, so that the technical problem of excessive prediction conclusion errors caused by large model generation illusion is solved, and the accuracy and credibility of event evolution prediction are improved.
Need to check novelty before this filing date? Find Prior Art