CN-121996760-A - Problem processing method and device based on artificial intelligence, computer equipment and medium

CN121996760ACN 121996760 ACN121996760 ACN 121996760ACN-121996760-A

Abstract

The application belongs to the technical field of artificial intelligence, and relates to a problem processing method, a device, computer equipment and a storage medium based on artificial intelligence, wherein the method comprises the steps of receiving problem data input by a user; the method comprises the steps of carrying out data generation processing on question data based on an enhancement generation strategy to obtain enhancement question data, carrying out combination processing on the question data and the enhancement question data to obtain combination data, carrying out reasoning processing on the combination data based on a target large language model to obtain an output result, carrying out consistency check on the output result to obtain a consistency check result, carrying out compliance detection and fact alignment check processing on the output result to obtain a compliance detection result and a fact check result, carrying out answer generation processing on the basis of the consistency check result, the compliance detection result and the fact check result to obtain answer data, and carrying out output processing on the answer data. The application can be applied to question-answer processing scenes in the field of finance and technology, and the stability and accuracy of generating answers are effectively improved through the application.

Inventors

WANG JIANZONG
ZHANG NAN
QU XIAOYANG

Assignees

平安科技（深圳）有限公司

Dates

Publication Date: 20260508
Application Date: 20260109

Claims (10)

1. The problem processing method based on artificial intelligence is characterized by comprising the following steps: receiving question data input by a user; performing data generation processing on the problem data based on a preset enhancement generation strategy to obtain corresponding enhancement problem data; combining the problem data with the enhanced problem data to obtain corresponding combined data; Carrying out reasoning processing on the combined data based on a preset target large language model to obtain an output result, and carrying out consistency verification on the output result to obtain a corresponding consistency verification result; Performing compliance detection and fact alignment verification processing on the output result to obtain a corresponding compliance detection result and a fact verification result; performing answer generation processing based on the consistency check result, the compliance detection result and the fact check result to obtain corresponding answer data; and outputting the answer data.
2. The method for processing the problem based on the artificial intelligence according to claim 1, wherein the step of performing data generation processing on the problem data based on a preset enhancement generation policy to obtain corresponding enhancement problem data specifically includes: Performing text disturbance enhancement processing on the problem data to obtain corresponding first problem data; Performing context package enhancement processing on the first problem data to obtain corresponding second problem data; performing cross-style repeating processing on the second problem data to obtain corresponding third problem data; And taking the third question data as the enhanced question data.
3. The method for processing the problem based on the artificial intelligence according to claim 1, wherein the step of performing inference processing on the combined data based on the preset target large language model to obtain an output result, and performing consistency check on the output result to obtain a corresponding consistency check result specifically comprises the steps of: invoking a pre-trained target large language model to perform parallel reasoning processing on the combined data to obtain a corresponding output result; Performing external output consistency check on the output result to obtain a corresponding first check result; performing internal representation consistency check on the output result to obtain a corresponding second check result; Integrating the first verification result and the second verification result to obtain a corresponding target verification result; And taking the target verification result as the consistency verification result.
4. The method for processing an artificial intelligence based question according to claim 1, wherein the step of generating answers based on the consistency check result, the compliance detection result and the fact check result to obtain corresponding answer data comprises: calling a preset rule knowledge base; Inquiring corresponding rule information from the rule knowledge base based on the problem data and the output result; Performing compliance detection on the output result and the rule information based on a preset regression classifier to obtain a corresponding compliance detection result; extracting key fact information from the output result; And carrying out fact alignment verification on the key fact information to obtain a corresponding fact verification result.
5. The method for processing an artificial intelligence based question according to claim 1, wherein the step of generating answers based on the consistency check result, the compliance detection result and the fact check result to obtain corresponding answer data comprises: Integrating the consistency check result, the compliance detection result and the fact check result to obtain corresponding integrated information; Performing logic analysis based on the integrated information to obtain a corresponding logic analysis result; screening a corresponding target answer template from a preset answer module library based on the logic analysis result; acquiring content information corresponding to the target answer template; Performing information filling processing on the target answer templates based on the content information to obtain filled target answer templates; And taking the filled target answer template as the answer data.
6. The method for processing questions based on artificial intelligence as claimed in claim 1, wherein the step of outputting the answer data comprises: constructing interpretation data corresponding to the answer data; optimizing the answer data based on a preset optimization strategy to obtain corresponding target answer data; Acquiring a preset output form; and carrying out output processing on the target answer data and the interpretation data based on the output form.
7. The method for processing questions based on artificial intelligence as claimed in claim 6, wherein the step of optimizing the answer data based on a preset optimization strategy to obtain corresponding target answer data comprises: optimizing the content richness of the answer data to obtain corresponding first answer data; Optimizing the content readability of the first answer data to obtain corresponding second answer data; carrying out semantic style optimization on the second answer data to obtain corresponding third answer data; performing scene adaptation optimization on the third answer data to obtain corresponding fourth answer data; and taking the fourth answer data as the target answer data.
8. An artificial intelligence based problem-handling device, comprising: The receiving module is used for receiving the problem data input by the user; The first generation module is used for carrying out data generation processing on the problem data based on a preset enhancement generation strategy to obtain corresponding enhancement problem data; the combination module is used for carrying out combination processing on the problem data and the enhanced problem data to obtain corresponding combination data; The processing module is used for carrying out reasoning processing on the combined data based on a preset target large language model to obtain an output result, and carrying out consistency check on the output result to obtain a corresponding consistency check result; The verification module is used for carrying out compliance detection and fact alignment verification processing on the output result to obtain a corresponding compliance detection result and a corresponding fact verification result; The second generation module is used for carrying out answer generation processing based on the consistency check result, the compliance detection result and the fact check result to obtain corresponding answer data; And the output module is used for carrying out output processing on the answer data.
9. A computer device comprising a memory having stored therein computer readable instructions which when executed implement the steps of the artificial intelligence based problem handling method of any of claims 1 to 7.
10. A computer readable storage medium having stored thereon computer readable instructions which when executed by a processor implement the steps of the artificial intelligence based problem handling method according to any of claims 1 to 7.

Description

Problem processing method and device based on artificial intelligence, computer equipment and medium Technical Field The application relates to the technical field of artificial intelligence, which can be applied to the field of financial science and technology, in particular to a problem processing method, a device, computer equipment and a storage medium based on artificial intelligence. Background In the traditional financial insurance service model, a Large Language Model (LLMs) has been applied to a number of tasks such as claim consultation, risk assessment, term interpretation and customer service, but there are significant problems in the problem handling process. In particular, current models are susceptible to catering bias, tending to generate replies by "good users" rather than strictly following factual logic. For example, when a customer challenges or misunderstands the terms of an insurance, the model may output a reply inconsistent with the terms of the contract by excessively catering to the emotion of the user, resulting in a lack of data stability and accuracy of the answer, which in turn may lead to disputes in claims or reduced trust of the customer. Taking the case of claim consultation in the financial field as an example, if a customer has misunderstanding on discipline terms (such as thinking that "non-professional sports injury" should belong to the claim range) when applying for claim settlement due to an accident, the traditional model may cater to the expectation of the user for quick claim settlement, misinterpret term ranges or fuzzy key limiting conditions, and finally lead to inconsistent claim settlement conclusion with contractual agreements. Such deviations not only compromise customer equity, but also increase the risk of operation and compliance costs of the insurance company. Therefore, there is a need to develop a problem processing method capable of improving the stability and accuracy of answer generation so as to improve the service reliability and reduce legal risks. Disclosure of Invention The embodiment of the application aims to provide a problem processing method, a device, computer equipment and a storage medium based on artificial intelligence, so as to solve the technical problem that the existing problem processing method based on a large language model has the problem that generated answers lack data stability and accuracy. In a first aspect, there is provided an artificial intelligence based problem handling method, comprising: receiving question data input by a user; performing data generation processing on the problem data based on a preset enhancement generation strategy to obtain corresponding enhancement problem data; combining the problem data with the enhanced problem data to obtain corresponding combined data; Carrying out reasoning processing on the combined data based on a preset target large language model to obtain an output result, and carrying out consistency verification on the output result to obtain a corresponding consistency verification result; Performing compliance detection and fact alignment verification processing on the output result to obtain a corresponding compliance detection result and a fact verification result; performing answer generation processing based on the consistency check result, the compliance detection result and the fact check result to obtain corresponding answer data; and outputting the answer data. In a second aspect, there is provided an artificial intelligence based problem handling apparatus comprising: The receiving module is used for receiving the problem data input by the user; The first generation module is used for carrying out data generation processing on the problem data based on a preset enhancement generation strategy to obtain corresponding enhancement problem data; the combination module is used for carrying out combination processing on the problem data and the enhanced problem data to obtain corresponding combination data; The processing module is used for carrying out reasoning processing on the combined data based on a preset target large language model to obtain an output result, and carrying out consistency check on the output result to obtain a corresponding consistency check result; The verification module is used for carrying out compliance detection and fact alignment verification processing on the output result to obtain a corresponding compliance detection result and a corresponding fact verification result; The second generation module is used for carrying out answer generation processing based on the consistency check result, the compliance detection result and the fact check result to obtain corresponding answer data; And the output module is used for carrying out output processing on the answer data. In a third aspect, a computer device is provided, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor implem