AI-driven financial knowledge management and retrieval system and method, storage medium and electronic equipment

Abstract

The invention provides an AI-driven financial knowledge management and retrieval system and method, a storage medium and electronic equipment, and the system adopts a layered architecture and comprises a knowledge processing layer which comprises an NLP analysis module, an entity recognition module and a knowledge annotation engine, the structured knowledge unit is used for converting an unstructured financial text into a structured knowledge unit with an associated three-dimensional label; the retrieval and analysis layer comprises a multi-dimensional query engine, a semantic matching module and a combined condition analyzer and is used for analyzing and executing complex query based on the three-dimensional label; and the interaction control layer comprises a natural language interface, an authority verification module and a risk assessment engine and is used for processing user interaction, authority control and security risk assessment. According to the method, efficient retrieval, automatic updating and safe sharing of knowledge can be achieved, and the defects of flexibility, precision and safety of a traditional system are overcome.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to an AI-driven financial knowledge management and retrieval system, method, storage medium, and electronic device. Background Technology

[0002] Traditional financial industry knowledge bases often employ a hierarchical directory-based management model, relying on manual classification and maintenance. Retrieval methods primarily rely on keyword matching, and common technologies include relational database-based storage systems and basic access control modules. However, existing technologies have significant shortcomings: First, access control is crude, with permission allocation dependent on manual settings, failing to dynamically adapt to knowledge sensitivity levels and easily leading to data leakage risks. Second, maintenance complexity is high, requiring multi-level approvals and manual operations for knowledge updates, resulting in low efficiency and a high risk of errors. Third, retrieval accuracy is insufficient, with keyword matching struggling to handle complex semantic queries involving multi-dimensional combinations of conditions. Fourth, sharing management is rigid, lacking intelligent risk assessment mechanisms and failing to balance the needs of knowledge sharing and privacy protection. Summary of the Invention

[0003] The technical problem to be solved by this invention is to provide an AI-driven financial knowledge management and retrieval system, method, storage medium and electronic device, which realizes efficient knowledge retrieval, automated updating and secure sharing, and solves the defects of traditional systems in terms of flexibility, accuracy and security.

[0004] This invention is implemented as follows: In a first aspect, the present invention provides an AI-driven financial knowledge management and retrieval system. The system adopts a layered architecture, including: a knowledge processing layer, including an NLP parsing module, an entity recognition module, and a knowledge annotation engine, for converting unstructured financial text into structured knowledge units with associated three-dimensional labels; The retrieval and analysis layer includes a multidimensional query engine, a semantic matching module, and a combined condition analyzer, used to parse and execute complex queries based on the three-dimensional labels; The interaction control layer, including a natural language interface, an access control module, and a risk assessment engine, is used to handle user interaction, access control, and security risk assessment.

[0005] Furthermore, the three-dimensional labels include customer dimensions, product dimensions, and employee dimensions.

[0006] Furthermore, the permission policy of the permission verification module is dynamically generated based on the sensitivity of the knowledge unit.

[0007] Furthermore, the risk assessment engine is used to perform real-time risk assessments on knowledge-sharing operations and dynamically adjust the accessibility of knowledge based on the assessment results.

[0008] Furthermore, in response to the update command received through the natural language interface, the system automatically parses and executes the update operation on the knowledge base after verification by the permission verification module.

[0009] Secondly, this invention provides an AI-driven method for financial knowledge management and retrieval, comprising: S1: Through the knowledge processing layer, the input unstructured financial text is parsed, entity is identified, and 3D annotation is performed to generate and store structured knowledge units; S2: Receives natural language queries or instructions from users through the natural language interface of the interactive control layer; S3: Through the retrieval and analysis layer, the semantics of the user query are parsed and transformed into combined query conditions based on three-dimensional labels; S4: Security control is implemented for query or operation requests through the permission verification module and risk assessment engine of the interaction control layer; S5: Returns query results or performs an update operation.

[0010] Furthermore, the three-dimensional annotation mentioned in step S1 refers to assigning labels to the three dimensions of customers, products, and employees.

[0011] Furthermore, the security controls in step S4 include verification based on dynamic permission policies and real-time risk assessment of sharing requests.

[0012] Thirdly, the present invention provides a computer-readable storage medium storing program instructions, specifically including the following steps: when the instructions are executed by a processor, the steps of implementing the AI-driven financial knowledge management and retrieval method described in the second aspect are implemented.

[0013] Fourthly, the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the program to implement the steps of the AI-driven financial knowledge management and retrieval method described in the second aspect.

[0014] This invention has the following advantages: By integrating artificial intelligence technology with three-dimensional structured data management, it constructs a knowledge management framework with "customer-product-employee" as the core dimension. Based on NLP, it achieves natural language interaction, combined with intelligent annotation, dynamic access control, and AI-driven shared governance, enabling efficient knowledge retrieval, automated updates, and secure sharing, thus overcoming the shortcomings of traditional systems in terms of flexibility, accuracy, and security. Attached Figure Description

[0015] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0016] Figure 1This is a schematic diagram of the system structure in Embodiment 1 of the present invention; Figure 2 This is a flowchart of the method in Embodiment 2 of the present invention; Figure 3 This is a schematic diagram of the structure of the storage medium in Embodiment 3 of the present invention; Figure 4 is a schematic diagram of the structure of the electronic device in Embodiment 4 of the present invention. Detailed Implementation

[0017] This invention provides an AI-driven financial knowledge management and retrieval system, method, storage medium, and electronic device to achieve efficient knowledge retrieval, automated updates, and secure sharing, thereby addressing the shortcomings of traditional systems in terms of flexibility, accuracy, and security. Example 1

[0018] This embodiment provides an AI-driven financial knowledge management and retrieval system, such as... Figure 1 As shown, a layered architecture is adopted, specifically including: The knowledge processing layer, including an NLP parsing module, an entity recognition module, and a knowledge annotation engine, is used to transform unstructured financial text into structured knowledge units with associated three-dimensional labels, where the three-dimensional labels include customer, product, and employee dimensions. The retrieval and analysis layer, including a multidimensional query engine, a semantic matching module, and a combined condition analyzer, is used to parse and execute complex queries based on three-dimensional labels; The interaction control layer, including a natural language interface, an access control module, and a risk assessment engine, is used to handle user interaction, access control, and security risk assessment. The access control policy of the access control module is dynamically generated based on the sensitivity of knowledge units. The risk assessment engine is used to perform real-time risk assessment on knowledge sharing operations and dynamically adjust the accessibility of knowledge based on the assessment results.

[0019] The system responds to update commands received through the natural language interface, and after verification by the permission verification module, automatically parses and executes the update operation on the knowledge base.

[0020] The system of this invention is deployed on a server cluster and a distributed storage device. The server cluster is responsible for carrying the training and inference tasks of all A1 modules in the knowledge processing layer, retrieval and analysis layer, and interaction control layer, utilizing its distributed computing capabilities to process massive amounts of data. The distributed storage device is used to store structured knowledge units, three-dimensional tag indexes, user permission policies, and system logs.

[0021] At the software level, the knowledge processing layer serves as the system's "intelligent entry point." When a new product manual PDF (unstructured text) is imported into the system, the NLP parsing module (which can be based on pre-trained models such as BERT) first performs deep semantic parsing. Subsequently, the entity recognition module (which can use sequence labeling models such as BiLSTM-CRF) accurately extracts "Customer A" (entity), "High-end Financial Product B" (entity), and their respective attributes from the text. The knowledge labeling engine then automatically tags this knowledge with three-dimensional labels: "Customer: A, Product: B, Source Employee: Manager C," and stores it as a structured knowledge unit in the distributed database.

[0022] When a user (such as a customer service representative) asks through the natural language interface of the interaction control layer (such as a chatbot window), "What is the latest rate for product B purchased by customer A?", the system's retrieval and analysis layer begins operation. The semantic matching module and the combined condition analyzer work together to parse this question into query conditions in the dimensions of "customer = A and product = B". The multidimensional query engine uses a pre-built three-dimensional index to quickly locate relevant knowledge points. Before returning the results, the permission verification module checks whether the customer service representative has permission to view the association information between customer A and product B. At the same time, the risk assessment engine may assess whether the query itself carries any risk. Finally, the system returns an accurate answer to the user. Example 2

[0023] This embodiment provides an AI-driven method for financial knowledge management and retrieval, such as... Figure 2 As shown, it includes: S1: Through the knowledge processing layer, the input unstructured financial text is parsed, entity is identified, and 3D labeled to generate and store structured knowledge units. The 3D labeling refers to assigning tags to customers, products, and employees in three dimensions. S2: Receives natural language queries or instructions from users through the natural language interface of the interactive control layer; S3: Through the retrieval and analysis layer, the semantics of the user query are parsed and transformed into combined query conditions based on three-dimensional labels; S4: Through the permission verification module and risk assessment engine of the interactive control layer, security control is performed on query or operation requests. Security control includes verification based on dynamic permission policies and real-time risk assessment of shared requests. S5: Returns query results or performs an update operation. Example 3

[0024] This embodiment provides a computer-readable storage medium, such as... Figure 3As shown, the program instructions are stored, including the following steps: when the instructions are executed by the processor, the steps of implementing an AI-driven financial knowledge management and retrieval method according to Embodiment 2 are implemented. Example 4

[0025] This embodiment provides an electronic device, such as... Figure 4 As shown, it includes a memory, a processor, and a computer program stored in the memory. When the processor executes the program, it implements the steps of an AI-driven financial knowledge management and retrieval method according to Embodiment 2.

[0026] The advantages of this invention lie in providing an AI-driven financial knowledge management and retrieval system, method, storage medium, and electronic device. By deeply integrating artificial intelligence technology with a three-dimensional structured data management model, it constructs a knowledge management system with "customer-product-employee" as its core dimension. This system not only supports efficient queries under complex conditions and significantly improves the accuracy of retrieval results, but also greatly reduces manual intervention in the knowledge update process, thereby effectively reducing system maintenance costs. Furthermore, this invention demonstrates excellent security; by optimizing data management processes, it significantly reduces the risk of data leakage, further enhancing system reliability. Simultaneously, its flexible knowledge sharing mechanism significantly improves the compliance sharing rate, providing a more efficient solution for internal and external collaboration within financial institutions.

[0027] While specific embodiments of the present invention have been described above, those skilled in the art should understand that the specific embodiments described are merely illustrative and not intended to limit the scope of the present invention. Equivalent modifications and variations made by those skilled in the art in accordance with the spirit of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. An AI-driven financial knowledge management and retrieval system, characterized in that: a hierarchical architecture is adopted, comprising: a knowledge processing layer, including an NLP analysis module, an entity recognition module and a knowledge labeling engine, for converting unstructured financial text into structured knowledge units with associated three-dimensional tags; a retrieval and analysis layer, including a multi-dimensional query engine, a semantic matching module and a combined condition analyzer, for parsing and executing complex queries based on the three-dimensional tags; an interactive control layer, including a natural language interface, a permission verification module and a risk assessment engine, for processing user interaction, permission control and security risk assessment. The three-dimensional tags include customer dimension, product dimension and employee dimension. The permission policy of the permission verification module is dynamically generated based on the sensitivity of the knowledge unit. The risk assessment engine is used for real-time risk assessment of knowledge sharing operations, and dynamically adjusts the accessibility of knowledge according to the assessment results. The system responds to update instructions received through the natural language interface, and automatically parses and executes update operations on the knowledge base after verification by the permission verification module. The method comprises the following steps: S1: parsing, entity recognition and three-dimensional labeling of input unstructured financial text through the knowledge processing layer, generating and storing structured knowledge units; S2: receiving user natural language queries or instructions through the natural language interface of the interactive control layer; S3: parsing the semantics of user queries and converting them into combined query conditions based on three-dimensional tags through the retrieval and analysis layer; S4: security control of query or operation request through the permission verification module and risk assessment engine of the interactive control layer; S5: returning query results or executing update operations. The three-dimensional labeling in step S1 refers to assigning labels to the three dimensions of customers, products and employees. The security control in step S4 includes dynamic permission policy-based verification and real-time risk assessment of sharing requests. When the instructions are executed by the processor, the steps of the AI-driven financial knowledge management and retrieval method according to claims 6-8 are implemented. The processor executes the program to implement the steps of the AI-driven financial knowledge management and retrieval method according to claims 6-8. ​ ​ ​ ​ 2. The AI-driven financial knowledge management and retrieval system of claim 1, wherein: ​ 3. The AI-driven financial knowledge management and retrieval system of claim 1, wherein: ​ 4. The AI-driven financial knowledge management and retrieval system of claim 1, wherein: ​ 5. The AI-driven financial knowledge management and retrieval system of claim 1, wherein: ​ 6. An AI-driven financial knowledge management and retrieval method applied to the AI-driven financial knowledge management and retrieval system of any one of claims 1 to 5, characterized in that: ​ ​ ​ ​ ​ ​ 7. The AI-driven financial knowledge management and retrieval method of claim 6, wherein: ​ 8. The AI-driven financial knowledge management and retrieval method of claim 6, wherein: ​ 9. A computer readable storage medium storing program instructions, characterized in that: ​ 10. An electronic device comprising a memory, a processor, and a computer program stored on the memory, characterized in that: ​

Images