Financial artificial intelligence platform and financial service execution method

Abstract

The invention provides a financial artificial intelligence platform and a financial service execution method, and the platform comprises a user interaction module which is used for obtaining user input and carrying out the preprocessing of the user input; the large language model processing module is used for performing intention recognition on the preprocessed user input based on a pre-trained large language model to obtain a user intention; and the Agent processing module is used for determining a financial service based on the user intention, dividing the financial service into a plurality of sub-tasks, determining a task execution plan of the plurality of sub-tasks based on a preset task priority and a task dependency relationship, and executing a financial operation corresponding to each sub-task based on the task execution plan. According to the invention, the flexibility and accuracy of financial service processing are improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a financial artificial intelligence platform and a method for executing financial services. Background Art

[0002] In the financial business processing scenario, there are some existing financial artificial intelligence platforms with large language models. Although they have achieved remarkable results in improving the efficiency and quality of financial services, there are still some shortcomings and defects. Summary of the invention

[0003] In view of this, an object of the present invention is to provide a financial artificial intelligence platform and a method for executing financial business, so as to improve the flexibility and accuracy of financial business processing.

[0004] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0005] In a first aspect, the present invention provides a financial artificial intelligence platform, including: a user interaction module, used to obtain user input and preprocess the user input; a large language model processing module, used to identify the intent of the preprocessed user input based on a pre-trained large language model to obtain the user intent; an Agent processing module, used to determine the financial business based on the user intent, and divide the financial business into multiple subtasks, and determine the task execution plan of the multiple subtasks based on pre-set task priorities and task dependencies, and execute the financial operation corresponding to each subtask based on the task execution plan.

[0006] Optionally, the Agent processing module is specifically used to: call the corresponding financial application or service interface based on the task execution plan to execute the financial operation corresponding to each subtask.

[0007] Optionally, it also includes: a feedback module, which is used to obtain the execution results of each subtask, and output the execution results to the user after processing; wherein the output form of the execution results includes one or more of the following forms: text form, voice broadcast, and chart display.

[0008] Optionally, it also includes: an optimization module, which is used to obtain user feedback information on the execution results, and optimize and update the large language model processing module and the Agent processing module based on the feedback information.

[0009] Optionally, it also includes: a monitoring module, which is used to: obtain in real time the operating status of the financial artificial intelligence platform during task execution, as well as log information during user interaction and task execution, and perform anomaly analysis based on the operating status and log information.

[0010] Optionally, the user interaction module is specifically used for: if the user input is text information, then performing word segmentation and keyword extraction on the text information; if the user input is voice, then performing voice recognition on the voice to obtain the corresponding text information, and performing word segmentation and keyword extraction on the text information; if the user input is an image or video, then extracting text information from the image or video, and performing word segmentation and keyword extraction on the text information.

[0011] Optionally, it also includes: a storage module for storing dialogue information, task execution information and business information of financial services during the user interaction process.

[0012] In a second aspect, the present invention provides a method for executing financial business, which is applied to any financial artificial intelligence platform provided in the first aspect, including: obtaining user input and preprocessing the user input; identifying the intent of the preprocessed user input based on a pre-trained large language model to obtain the user intent; determining the financial business based on the user intent, and dividing the financial business into multiple subtasks; determining the task execution plan of multiple subtasks based on pre-set task priorities and task dependencies, and executing the financial operations corresponding to each subtask based on the task execution plan.

[0013] In a third aspect, the present invention provides an electronic device comprising a processor and a memory, wherein the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the steps of the method provided in the second aspect above.

[0014] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the method provided in the second aspect are executed.

[0015] The present invention brings the following beneficial effects:

[0016] The financial artificial intelligence platform and the execution method of financial business provided by the present invention include: a user interaction module for obtaining user input and preprocessing the user input; a large language model processing module for performing intent recognition on the preprocessed user input based on a pre-trained large language model to obtain user intent; an agent processing module for determining financial business based on user intent, dividing the financial business into multiple subtasks, and determining the task execution plan of multiple subtasks based on pre-set task priorities and task dependencies, and executing the financial operations corresponding to each subtask based on the task execution plan. The above platform first recognizes the intent of the user input through the large language model processing module, and then determines the financial business required by the user according to the user intent through the agent processing module, and splits the financial business into multiple subtasks, determines the task execution plan according to the task priority and task dependency, and then executes the financial operation corresponding to each subtask. The present invention determines the financial task through the user intent, and splits the financial business required by the user into multiple subtasks for execution, thereby improving the flexibility and accuracy of financial business processing.

[0017] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description, claims and drawings.

[0018] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the preferred embodiments are specifically mentioned below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0020] Figure 1 A schematic diagram of the structure of a financial artificial intelligence platform provided by an embodiment of the present invention;

[0021] Figure 2 A schematic diagram of the structure of another financial artificial intelligence platform provided by an embodiment of the present invention;

[0022] Figure 3 A flowchart of a method for executing a financial service provided by an embodiment of the present invention;

[0023] Figure 4A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] At present, in the financial business processing scenarios, there are some financial artificial intelligence platforms with large language models. Although they have achieved remarkable results in improving the efficiency and quality of financial services, they still have certain defects in data processing, intelligent decision-making, task processing, etc.

[0026] Based on this, the embodiments of the present invention provide a financial artificial intelligence platform and a method for executing financial business, which can improve the flexibility and accuracy of financial business processing.

[0027] To facilitate understanding of this embodiment, a financial artificial intelligence platform disclosed in an embodiment of the present invention is first introduced in detail. Before using the financial artificial intelligence platform (i.e., Agent system) to process financial business, the platform needs to be initialized and configured, mainly including: firstly, the platform is initialized, including: loading necessary system libraries, library files, configuration parameters and dependencies, such as: deep learning framework, natural language processing library, etc.; then, computing resources (such as CPU, GPU, memory) are allocated to the language model and Agent components; then, the pre-trained large language model is loaded from the storage medium (such as: hard disk, cloud storage) into the memory, and the large language model includes: GPT series, BERT series, etc., to prepare for subsequent reasoning and generation tasks. Loading a pre-trained large language model usually involves reading the model parameters from the model storage location and loading them onto a computing device (such as a GPU), initializing the model parameters, including the maximum sequence length, batch size, etc., to ensure that the model is in a usable state; finally, configuring the relevant parameters of the Agent, such as the communication protocol, interface address, task execution strategy, task manager, knowledge base, etc., setting the interaction interface between the Agent and other components of the financial application platform, API calls, and database connections. These configurations are usually adjusted according to the needs of the actual application scenario.

[0028] See also Figure 1 The structure diagram of a financial artificial intelligence platform shown in FIG. 1 shows that the platform mainly includes the following parts:

[0029] The user interaction module 101 is used to obtain user input and pre-process the user input.

[0030] In one implementation, user input may be received through a communication interface (such as a Web front end, a mobile application, etc.), and the user input may be natural language text, voice, image, video, or other forms of data.

[0031] Preprocess user input, including noise removal, word segmentation, part-of-speech tagging, speech recognition to text, etc., and use natural language processing technology to parse user input and extract key information for subsequent natural language understanding and processing.

[0032] In specific implementation, if the user input is text information, the text information is segmented and keywords are extracted; if the user input is voice, voice recognition is performed on the voice to obtain the corresponding text information, and the text information is segmented and keywords are extracted; if the user input is an image or video, the text information is extracted from the image or video, and the text information is segmented and keywords are extracted.

[0033] The large language model processing module 102 is used to perform intent recognition on the pre-processed user input based on the pre-trained large language model to obtain the user intent.

[0034] In one implementation, a pre-trained large language model is used to understand and parse the user input to identify the user's intention and needs, which generally involves natural language understanding (NLU) technology, such as intent recognition, entity extraction, etc. For example, a pre-trained large language model is used to understand the key information of the user input, such as the type, amount, time limit, etc. of the financial product, and based on the parsing results, the user's ultimate purpose or need (i.e., user intention) is identified, such as purchasing financial products, checking account balances, transferring money, etc.

[0035] In specific implementation, this module adopts advanced large language model technology, such as pre-trained models based on Transformer architecture, and has powerful text generation and understanding capabilities through large-scale training data. Specifically, this module can include a shared network and multiple domain expert networks, and select appropriate domain expert networks through the router module to process the corresponding input features, thereby improving the accuracy of large language models in processing texts in various fields. In addition, this module can also use knowledge activation networks (KANs) based on text data in the knowledge base to update the first large language model to obtain the second large language model, and perform multi-task training on the second large language model based on the text preference data set to obtain the target large language model, so as to improve the flexibility and expression ability of the large language model. At the same time, this module also has domain adaptation capabilities, and can fine-tune the model according to the specific needs of the financial field to improve the processing effect of the model on financial text data.

[0036] The agent processing module 103 is used to determine the financial business based on the user's intention, divide the financial business into multiple subtasks, determine the task execution plan of the multiple subtasks based on the pre-set task priorities and task dependencies, and execute the financial operation corresponding to each subtask based on the task execution plan.

[0037] In one embodiment, the Agent processing module 103 performs task planning and execution through the Agent system. The module uses advanced Agent technology, based on the output results of the large language model module, combined with its own knowledge base and rule base, to make intelligent decisions and actions, and build an intelligent Agent that can simulate human behavior. These Agents can interact with users intelligently, understand user intentions, and perform corresponding operations according to user needs. At the same time, the Agent module also has autonomous learning capabilities and can continuously optimize its own behavior strategies based on user feedback and behavior. The Agent module can also collaborate and communicate with other Agents to achieve more complex financial tasks.

[0038] Specifically, the user intention is first converted into a specific financial service. For example, if the user wants to "view last month's transaction records", the financial service is defined as querying account activities within a specified time period. Then, according to the user's intention, the financial service is decomposed into multiple subtasks or steps, and a strategy and plan for task execution (i.e., task execution plan) is formulated, including: selecting appropriate financial applications or services, determining the execution order, etc. Specifically, when formulating a task execution plan, it is necessary to consider a variety of factors, such as: task priority, resource availability, user preference, task dependency, etc. In this embodiment, a task execution plan is formulated based on the priority and dependency of subtasks. Finally, according to the task execution plan, the corresponding financial application or service interface is called to trigger different financial applications or services (including: account management, transaction processing, risk assessment and other financial services), such as: sending HTTP requests, parsing response results, etc., passing necessary parameters and data, and performing specific financial operations.

[0039] Taking "view last month's transaction records" as an example, subtasks may include verifying user identity, connecting to bank APIs, requesting transaction data, filtering and displaying data, etc. Then determine the execution order and priority of each subtask, and design the most appropriate execution strategy based on the task characteristics and system capabilities. For example, for operations involving sensitive data, security measures should be prioritized. And formulate detailed execution steps for each subtask, including the expected time frame, required resources and technical means. Finally, execute the task according to the established strategy and plan, and monitor the progress of the task in real time.

[0040] The financial artificial intelligence platform provided by the present invention first recognizes the intention of the user input through the large language model processing module, and then determines the financial business required by the user according to the user intention through the Agent processing module, and splits the financial business into multiple subtasks, determines the task execution plan according to the task priority and task dependency, and then executes the financial operation corresponding to each subtask. The present invention determines the financial task according to the user intention, and splits the financial business required by the user into multiple subtasks for execution, thereby improving the flexibility and accuracy of financial business processing.

[0041] See also Figure 2 As shown, the above platform also includes: a feedback module 104, which is used to: obtain the execution result of each subtask, and output the execution result to the user after processing; wherein the output form of the execution result includes but is not limited to the following forms: text form, voice broadcast, and chart display.

[0042] In one embodiment, the execution result of the financial task is obtained, and the execution result of the financial task is processed and formatted, such as data conversion, aggregation, visualization, etc., so that the user can understand and use it. The processed execution result is presented to the user in natural language or other forms, and the display method includes but is not limited to: text output, voice broadcast, chart display, etc., and necessary explanations and instructions are provided to help the user understand the results.

[0043] See also Figure 2 As shown, the above platform also includes: an optimization module 105, which is used to: obtain user feedback information on the execution result, and optimize and update the large language model processing module and the Agent processing module based on the feedback information.

[0044] In one embodiment, the large language model processing module and the agent processing module can be iterated and optimized based on user feedback and needs. For example, if the user has questions about the execution results or needs more detailed information, the agent can further explain or provide more data to continuously improve the performance and user experience of the agent. At the same time, self-learning and optimization can be performed based on feedback information to improve the accuracy and efficiency of future decisions.

[0045] See also Figure 2 As shown, the above platform also includes: a monitoring module 106, which is used to: obtain in real time the operating status of the financial artificial intelligence platform during task execution, as well as log information during user interaction and task execution, and perform abnormal analysis based on the operating status and log information.

[0046] During the specific implementation, the operating status of the platform can be monitored in real time, including: computing resource usage, response time, task execution efficiency, etc. When an abnormality occurs during operation, an alarm is triggered in time to handle system failures or abnormalities. When necessary, the fault recovery mechanism is started to restore the normal operation of the system, which helps to discover and solve potential problems in a timely manner.

[0047] In addition, it can also record key information and log information in the process of user interaction and task execution, and discover potential problems and improvement points through log analysis. At the same time, according to the needs of actual application scenarios and the development of technology, it regularly updates and upgrades the large language model and related components of the Agent, adds new functions, optimizes performance, fixes security vulnerabilities, etc., introduces new financial knowledge and data, and improves the performance, generalization ability and accuracy of the platform.

[0048] See also Figure 2 As shown, the above platform also includes: a storage module 107, which is used to store dialogue information, task execution information and business information of financial services during the user interaction process.

[0049] In specific implementation, the platform's knowledge base and memory information can be stored and managed, including user preferences, historical conversation information, task execution status, etc. This information can be used to support the understanding, reasoning, and generation capabilities of large language models.

[0050] Specifically, the agent can maintain a short-term memory to store conversation context information, which helps to maintain contextual consistency in multiple rounds of dialogue. It can also maintain a long-term memory to store long-term information such as user characteristics and business data, and continuously update the long-term memory based on user feedback information and new business data. The agent continuously interacts and learns with users, and optimizes its ability to understand and execute tasks by fine-tuning the model and improving the strategy.

[0051] Furthermore, it also includes: a user interface module, which provides users with a friendly interactive interface and supports users to interact with the platform through natural language. Users can enter queries or instructions through the interface to communicate and exchange with the platform, and view the output results of the large language model processing module and the Agent processing module to obtain the required financial information and services. The user interface module can also optimize and improve the platform based on user feedback and needs. At the same time, the interface also supports a variety of visualization tools, such as charts, reports, etc., to help users better understand the analysis results.

[0052] The above-mentioned financial artificial intelligence platform provided by the embodiment of the present invention can be applied to multiple scenarios, such as: investment strategy formulation, risk management, customer service, etc. In the investment strategy formulation scenario, the Agent can formulate personalized investment strategies based on market trends and user needs, and provide investment advice and decision support to users by analyzing historical data and predicting future market trends. In the risk management scenario, the Agent can evaluate loan risks, market risks, etc., and propose corresponding risk control measures, discover potential risk points through real-time monitoring and analysis of data, and conduct early warning and intervention in a timely manner; in the customer service scenario, the Agent can answer users' consulting questions, provide personalized service experience, communicate with users through natural language interaction, and provide corresponding solutions and suggestions based on user needs.

[0053] The Agent system involved in this embodiment is introduced in detail below.

[0054] Agent systems usually include four core modules (or components): planning, memory, tools, and action. These modules work together to enable agents to analyze situations, make decisions, and perform tasks autonomously.

[0055] The planning module is responsible for formulating long-term and short-term action plans based on current goals and environmental conditions, including planning with uncertain factors and how to effectively achieve the set goals, so that the agent can act in an organized manner in a complex and dynamic environment.

[0056] The memory module is divided into short-term memory and long-term memory. Short-term memory focuses on immediate status perception and supports immediate decision-making; long-term memory stores information through databases or special memory systems for subsequent retrieval and utilization, providing information support for subsequent actions.

[0057] The Tools module is used to process data, analyze data, make inferences, and make decisions. By interconnecting the model with the outside world, the model can not only perceive the world, but also change the external state by using tools. In the financial field, the model can be given the ability to perceive real-time changes in the financial market, such as calling query APIs to obtain fund data, transaction data, etc.

[0058] The Action module performs specific operations based on the plan and the current environment, which is a direct reflection of the Agent's interaction with the outside world. By calling appropriate tools, the Agent can achieve its set goals and demonstrate its actual ability to solve problems.

[0059] The specific implementation steps of the Agent system include:

[0060] 1. Environmental perception and data collection:

[0061] Agents collect environmental information through sensors or data interfaces, including financial market data, user behavior data, etc. These data are used to build the knowledge base and decision-making basis of the agent. Agents support multiple interaction methods, including natural language, images, videos, etc., so that agents can present complex data and information more intuitively and improve user experience.

[0062] 2. Data analysis and decision making:

[0063] The agent uses a large language model to analyze and interpret the collected data, and combines the knowledge base and rule base to formulate corresponding action plans or decisions.

[0064] 3. Task planning and decomposition:

[0065] The agent breaks down complex tasks into multiple subtasks for better management and execution. At the same time, the planning module ensures that each subtask is achievable and helps achieve the final goal.

[0066] 4. Tool call and execution:

[0067] According to the task requirements, the Agent calls the corresponding tools or APIs to perform specific operations. These tools can be data processing tools, query tools, transaction execution tools, etc.

[0068] 5. Result feedback and learning optimization:

[0069] After the agent performs the operation, it collects the results and provides feedback. Based on the feedback results, the agent conducts self-learning and optimization to improve the accuracy and efficiency of future decisions.

[0070] The above-mentioned financial artificial intelligence platform provided by the embodiment of the present invention improves the intelligence level of the platform, so that the platform can understand user intentions more accurately and provide more accurate and personalized financial services; enhances the adaptive ability of the platform, so that the platform can continuously maintain efficient and accurate financial service levels to meet the ever-changing needs of users; improves the user's interactive experience, so that users can use financial services more conveniently and intuitively, and improve user satisfaction and loyalty; improves the processing efficiency of the platform, so that the platform can respond to user requests more quickly and provide more timely and accurate financial services; improves the operating efficiency of the platform, reduces the operating costs of financial institutions, and enables financial institutions to focus more on the development and innovation of core businesses.

[0071] For the financial artificial intelligence platform provided in the above embodiments, the present invention also provides a method for executing financial services applied to the platform, see Figure 3 The figure shows a method for executing a financial service, which mainly includes the following steps S301 to S304:

[0072] Step S301: obtaining user input and preprocessing the user input;

[0073] Step S302: performing intent recognition on the preprocessed user input based on the pre-trained large language model to obtain the user intent;

[0074] Step S303: determining a financial service based on the user's intention, and dividing the financial service into a plurality of subtasks;

[0075] Step S304: Based on the preset task priorities and task dependencies, determine the task execution plans of the multiple subtasks, and execute the financial operations corresponding to each subtask based on the task execution plans.

[0076] The method for executing the above-mentioned financial business provided by the present invention first recognizes the intention of the user input through the large language model processing module, and then determines the financial business required by the user according to the user intention through the Agent processing module, and splits the financial business into multiple subtasks, determines the task execution plan according to the task priority and task dependency, and then executes the financial operation corresponding to each subtask. The present invention determines the financial task according to the user intention, and splits the financial business required by the user into multiple subtasks for execution, thereby improving the flexibility and accuracy of financial business processing.

[0077] It should be noted that the implementation principle and technical effects of the method provided in the embodiment of the present invention are the same as those of the aforementioned platform embodiment. For the sake of brief description, for matters not mentioned in the method embodiment, reference may be made to the corresponding contents in the aforementioned platform embodiment.

[0078] An embodiment of the present invention further provides an electronic device. Specifically, the electronic device includes a processor and a storage device. The storage device stores a computer program, and when the computer program is executed by the processor, it executes the method described in any one of the above implementation methods.

[0079] Figure 4 A structural diagram of an electronic device provided in an embodiment of the present invention, the electronic device 100 includes: a processor 40, a memory 41, a bus 42 and a communication interface 43, wherein the processor 40, the communication interface 43 and the memory 41 are connected via the bus 42; the processor 40 is used to execute an executable module stored in the memory 41, such as a computer program.

[0080] The memory 41 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 43 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used.

[0081] The bus 42 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0082] Among them, the memory 41 is used to store programs, and the processor 40 executes the program after receiving the execution instruction. The method executed by the device for flow process definition disclosed in any embodiment of the above-mentioned embodiment of the present invention can be applied to the processor 40 or implemented by the processor 40.

[0083] The processor 40 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 40. The above processor 40 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present invention can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The steps of the method disclosed in the embodiments of the present invention can be directly embodied as a hardware decoding processor to execute, or can be executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 41, and the processor 40 reads the information in the memory 41 and completes the steps of the above method in combination with its hardware.

[0084] The computer program product of the readable storage medium provided in the embodiment of the present invention includes a computer-readable storage medium storing program code, and the instructions included in the program code can be used to execute the methods described in the previous method embodiments. The specific implementation can be referred to the previous method embodiments, which will not be repeated here.

[0085] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0086] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A financial artificial intelligence platform, characterized in that: include: A user interaction module, used to obtain user input and pre-process the user input; A large language model processing module is used to identify the intent of the pre-processed user input based on the pre-trained large language model to obtain the user intent; The agent processing module is used to determine the financial business based on the user intention, divide the financial business into multiple subtasks, and determine the task execution plan of the multiple subtasks based on pre-set task priorities and task dependencies, and execute the financial operation corresponding to each subtask based on the task execution plan.

2. The platform according to claim 1, characterized in that The Agent processing module is specifically used for: Based on the task execution plan, the corresponding financial application or service interface is called to execute the financial operation corresponding to each subtask.

3. The platform according to claim 1, characterized in that: Also includes: The feedback module is used to obtain the execution result of each subtask, and output the execution result to the user after processing; wherein the output form of the execution result includes one or more of the following forms: text form, voice broadcast, and chart display.

4. The platform according to claim 3, characterized in that: Also includes: The optimization module is used to obtain user feedback information on the execution result, and optimize and update the large language model processing module and the Agent processing module based on the feedback information.

5. The platform according to claim 1, characterized in that: Also includes: The monitoring module is used to: obtain in real time the operating status of the financial artificial intelligence platform during task execution, as well as log information during user interaction and task execution, and perform abnormal analysis based on the operating status and log information.

6. The platform according to claim 1, characterized in that: The user interaction module is specifically used for: If the user input is text information, performing word segmentation and keyword extraction on the text information; If the user input is voice, voice recognition is performed on the voice to obtain corresponding text information, and word segmentation and keyword extraction are performed on the text information; If the user input is an image or a video, text information is extracted from the image or the video, and word segmentation and keyword extraction are performed on the text information.

7. The platform according to claim 1, characterized in that Also includes: The storage module is used to store the dialogue information, task execution information and business information of the financial business during the user interaction process.

8. A method for executing financial business, characterized in that: The financial artificial intelligence platform applied to any one of claims 1 to 7 comprises: Obtaining user input and preprocessing the user input; Based on the pre-trained large language model, the pre-processed user input is used to identify the user's intent and obtain the user's intent. Determine a financial service based on the user intention, and divide the financial service into a plurality of subtasks; Based on the preset task priorities and task dependencies, a task execution plan for the plurality of subtasks is determined, and the financial operation corresponding to each subtask is executed based on the task execution plan.

9. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores computer executable instructions that can be executed by the processor, and the processor executes the computer executable instructions to implement the steps of the method according to claim 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to claim 8 are performed.

Images