Automated process generation method, intelligent agent and automated process generation system

Abstract

The present application discloses an automated process generation method, an intelligent agent, and an automated process generation system, which relate to the field of robotic process automation technology. The method comprises: monitoring the natural language input by the user; when the natural language describing the target task is monitored, summarizing and arranging it to obtain the key information of the target task; according to the automated process component knowledge base, filtering the components matching the operation steps contained in the key information of the target task from the automated process component library to obtain a matching component set and obtain matching component description information; using a large language model, based on the key information of the target task and the matching component description information, obtaining automated process generation guidance information; according to the automated process generation guidance information, assembling each component in the matching component set to obtain an automated process. The automated process generation method disclosed in the present application can automatically and efficiently generate an automated process according to the natural language describing the target task input by the user.

Description

Technical Field

[0001] The present application relates to the field of robotic process automation technology, and in particular to an automated process generation method, an intelligent agent, and an automated process generation system. Background Art

[0002] Robotic Process Automation (RPA) is an automation technology that uses specific "robot software" to perform repetitive, low-value and regular tasks (such as data entry, information organization, email reply, etc.) to replace manual operations. RPA technology improves productivity, reduces error rates, enhances consistency, and enables enterprises to respond to changes more quickly, thereby improving efficiency and reducing costs.

[0003] The current automated process generation solution is: the user first observes the human operation process, and then uses the orchestration tool to design the automated process based on the human operation process. Although the current automated process generation solution can generate automated processes, the efficiency of automated process generation is low. Summary of the invention

[0004] In view of this, the present application provides an automated process generation method, an intelligent agent, and an automated process generation system, which are used to solve the problem of low generation efficiency of existing automated process generation solutions. The technical solution is as follows:

[0005] The first aspect of the present application provides an automated process generation method, comprising:

[0006] Monitor the natural language input by the user;

[0007] When natural language describing the target task input by the user is monitored, the natural language describing the target task is summarized and sorted to obtain key information of the target task, wherein the key information of the target task includes task content, task objectives and operation steps;

[0008] According to the automation process component knowledge base, components matching the operation steps are screened from the automation process component library to obtain a matching component set, and matching component description information is obtained, wherein the automation process component knowledge base includes description information of each component in the automation process component library, and the matching component description information includes description information of each component in the matching component set;

[0009] Using a large language model, based on the target task key information and the matching component description information, obtaining automated process generation guidance information, wherein the automated process generation guidance information includes component assembly information, and the component assembly information is used to indicate how to assemble components in the matching component set;

[0010] According to the component assembly information, each component in the matching component set is assembled to obtain an automated process.

[0011] In a possible implementation, the natural language describing the target task is summarized and sorted to obtain key information of the target task, including:

[0012] Using a fine-tuned finite-scale language model, the natural language describing the target task is summarized and organized to obtain key information of the target task;

[0013] The fine-tuned finite-scale language model is obtained by fine-tuning a pre-trained finite-scale language model using training data corresponding to a plurality of automation tasks, wherein the training data corresponding to an automation task includes a natural language describing the automation task and corresponding key task information.

[0014] In a possible implementation, the target task key information and the matching component description information are both structured natural languages;

[0015] The using of the large language model, based on the target task key information and the matching component description information, to obtain the automatic process generation guidance information includes:

[0016] Organizing the target task key information, the matching component description information and the preset instruction prompt information into a prompt instruction, wherein the prompt instruction is used to prompt the large language model to generate an automated process to generate guidance information according to the target task key information and the matching component description information;

[0017] The prompt instruction is input into a large language model to obtain automated process generation guidance information output by the large language model.

[0018] In a possible implementation, the component assembly information is structured code data;

[0019] The step of assembling each component in the matching component set according to the component assembly information to obtain an automated process includes:

[0020] Determining the assembly order of each component in the matching component set and the parameter configuration of each component by parsing the structured code data;

[0021] According to the assembly order of each component in the matching component set and the parameter configuration of each component, and in combination with the matching component description information, each component in the matching component set is assembled to obtain an automated process.

[0022] In a possible implementation, the description information of any component included in the automation process component library includes: a component name of the component and a component function description of the component;

[0023] The component function description of the component includes: the specific function of the component, the input parameters of the component, the output parameters of the component and the usage scenario of the component.

[0024] In a possible implementation, the automated process generating guidance information further includes: the user operation intention inferred by the large language model according to the target task key information and the matching component description information;

[0025] The automated process generation method further comprises:

[0026] outputting the automated process;

[0027] When an adjustment suggestion for the automation process input by the user is monitored, the automation process generation guidance information is reacquired based on the adjustment suggestion and the automation process generation guidance information acquired last time by using the large language model;

[0028] Component assembly information included in the re-acquired automated process generation guidance information is used to assemble components to obtain a new automated process.

[0029] In a possible implementation, the automated process generation method further includes:

[0030] Check if there are any components missing parameters in the output automation process;

[0031] If there are components with missing parameters in the output automation process, natural language describing the missing parameters is output so that the user can supplement the parameters, wherein the natural language describing the missing parameters is used to indicate the components with missing parameters, the parameters missing from the components with missing parameters, and the impact of the missing parameters on the automation process.

[0032] In a possible implementation, the automated process generation method further includes:

[0033] If the adjustment suggestion includes a suggestion for a missing component, determining the possibly missing component from the automation process component library according to the adjustment suggestion, and obtaining description information of the possibly missing component;

[0034] The using the large language model to reacquire the automated process generation guidance information based on the adjustment suggestion and the automated process generation guidance information acquired last time includes:

[0035] The large language model is used to reacquire the automated process generation guidance information based on the adjustment suggestions, the description information of the possibly missing components, and the automated process generation guidance information acquired last time.

[0036] The second aspect of the present application provides an intelligent agent, including: an input monitoring module, a task description processing module, a component matching module, an automated process generation guidance information acquisition module, and a component assembly module;

[0037] The input monitoring module is used to monitor the natural language input by the user;

[0038] The task description processing module is used to summarize and organize the natural language describing the target task when the input monitoring module monitors the natural language describing the target task input by the user to obtain key information of the target task, wherein the key information of the target task includes task content, task objectives and operation steps;

[0039] The component matching module is used to screen the components matching the operation steps from the automation process component library according to the automation process component knowledge base to obtain a matching component set and obtain matching component description information, wherein the automation process component knowledge base contains description information of each component in the automation process component library, and the matching component description information includes description information of each component in the matching component set;

[0040] The automated process generation guidance information acquisition module is used to use the large language model to acquire automated process generation guidance information based on the target task key information and the matching component description information, wherein the automated process generation guidance information includes component assembly information, and the component assembly information is used to indicate how to assemble components in the matching component set;

[0041] The component assembly module is used to assemble each component in the matching component set according to the component assembly information to obtain an automated process.

[0042] A third aspect of the present application provides an automated process generation system, including: an agent and a large language model module;

[0043] The intelligent agent is used to listen to the natural language input by the user, and when listening to the natural language input by the user describing the target task, the intelligent agent summarizes and organizes the natural language describing the target task to obtain key information of the target task, wherein the key information of the target task includes task content, task objectives and operation steps, and according to the automated process component knowledge base, the intelligent agent screens the components matching the operation steps from the automated process component library to obtain a matching component set, and obtains matching component description information, wherein the automated process component knowledge base contains description information of each component in the automated process component library, and the matching component description information includes description information of each component in the matching component set;

[0044] The large language model module is used to generate automatic process generation guidance information based on the target task key information and the matching component description information, wherein the automatic process generation guidance information includes component assembly information, and the component assembly information is used to indicate how to assemble components in the matching component set;

[0045] The intelligent agent is also used to assemble each component in the matching component set according to the component assembly information to obtain an automated process.

[0046] By means of the above technical scheme, the automated process generation method provided by the present application can monitor the natural language input by the user. When the natural language describing the target task input by the user is monitored, the natural language describing the target task is first summarized and sorted to obtain the key information of the target task, and then the components matching the operation steps contained in the key information of the target task are screened from the automated process component library according to the automated process component knowledge base to obtain a matching component set and obtain the matching component description information, and then the large language model is used to obtain the automated process generation guidance information based on the key information of the target task and the matching component description information, and finally, according to the component assembly information contained in the automated process generation guidance information, each component in the matching component set is assembled to obtain an automated process. The automated process generation method provided by the present application can automatically generate an automated process according to the natural language describing the target task input by the user. Compared with the scheme of manually arranging and designing the automated process, the generation efficiency of the automated process is greatly improved. In addition, since the automated process generation method provided by the present application does not require manual arrangement and design, the influence of human subjective factors on the generation of the automated process is avoided. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0048] Figure 1 A schematic diagram of a system architecture involved in this application;

[0049] Figure 2 A schematic diagram of a hardware structure of a terminal provided in an embodiment of the present application;

[0050] Figure 3 A schematic diagram of a hardware structure of a server provided in an embodiment of the present application;

[0051] Figure 4 A flowchart of an automated process generation method provided in an embodiment of the present application;

[0052] Figure 5 A flowchart of another automated process generation method provided in an embodiment of the present application;

[0053] Figure 6 A schematic diagram of the structure of an intelligent agent provided in an embodiment of the present application;

[0054] Figure 7 A schematic diagram of the structure of the automated process generation system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0055] The following describes the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. The terms used in the implementation method section of the present application are only used to explain the specific embodiments of the present application, and are not intended to limit the present application.

[0056] The embodiments of the present application are described below in conjunction with the accompanying drawings. Those skilled in the art will appreciate that, with the development of technology and the emergence of new scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0057] The terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and need not be used to describe a specific order or sequential order. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances, which is only to describe the distinction mode adopted by the objects of the same attributes when describing in the embodiments of the present application. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, so that the process, method, system, product or equipment comprising a series of units need not be limited to those units, but may include other units that are not clearly listed or inherent to these processes, methods, products or equipment.

[0058] Before introducing the automated process generation method provided by this application, the system architecture involved in this application is first explained.

[0059] In one possible implementation, Figure 1 As shown, the system architecture involved in the present application may include a terminal 101 and a server 102, and the terminal 101 may interact with the server 102 via a network (wired network or wireless network). The server 102 may include one or more servers ( Figure 1 A server is included as an example for explanation).

[0060] Next, the product form of the terminal 101 will be described.

[0061] The above-mentioned terminal 101 can be a mobile phone, a tablet computer, a wearable device, a vehicle-mounted device, an augmented reality (AR) / virtual reality (VR) device, a laptop computer, an ultra-mobile personal computer (UMPC), a netbook, a personal digital assistant (PDA), etc., and the embodiments of the present application do not impose any restrictions on this.

[0062] Figure 2 An optional hardware structure diagram of the terminal 101 is shown.

[0063] refer to Figure 2 As shown, the terminal 101 may include a radio frequency unit 210, a memory 220, an input unit 230, a display unit 240, a camera 250 (optional), an audio circuit 260 (optional), a speaker 261 (optional), a microphone 262 (optional), an earphone jack 263 (optional), a processor 270, an external interface 280, a power supply 290 and other components. Those skilled in the art will appreciate that Figure 2The terminal 101 is merely an example and does not constitute a limitation on the terminal 101 , which may include more or less components than those shown in the figure, or a combination of certain components, or different components.

[0064] The input unit 230 can be used to receive input digital or character information, and generate key signal input related to the user settings and function control of the terminal 101. Specifically, the input unit 230 may include a touch screen 231 (optional) and / or other input devices 232. The touch screen 231 can collect the user's touch operations on or near it (such as the user's operation on or near the touch screen using any suitable object such as fingers, joints, stylus, etc.), and drive the corresponding connection device according to a pre-set program. The touch screen can detect the user's touch action on the touch screen, convert the touch action into a touch signal and send it to the processor 270, and can receive and execute the command sent by the processor 270; the touch signal at least includes the touch point coordinate information. The touch screen 231 can provide an input interface and an output interface between the terminal 101 and the user. In addition, the touch screen can be implemented using multiple types such as resistive, capacitive, infrared, and surface acoustic waves. In addition to the touch screen 231, the input unit 230 can also include other input devices. Specifically, other input devices 232 may include, but are not limited to, one or more of a physical keyboard, function keys (such as volume control keys, switch keys, etc.), a trackball, a mouse, a joystick, and the like.

[0065] The display unit 240 may be used to display information input by a user or information provided to a user, various menus of the terminal 101, an interactive interface, file display, and / or playback of any multimedia file.

[0066] The memory 220 can be used to store instructions and data. The memory 220 can mainly include an instruction storage area and a data storage area. The data storage area can store various data, such as multimedia files, texts, etc. The instruction storage area can store software units such as operating systems, applications, and instructions required for at least one function, or their subsets and extensions. It can also include a non-volatile random access memory; it provides the processor 270 with hardware, software, and data resources including management of computing and processing equipment, and supports control software and applications. It is also used for the storage of multimedia files, and the storage of running programs and applications.

[0067] The processor 270 is the control center of the terminal 101. It uses various interfaces and lines to connect various parts of the entire terminal 101. By running or executing instructions stored in the memory 220 and calling data stored in the memory 220, it executes various functions of the terminal 101 and processes data, thereby controlling the terminal 101 as a whole. Optionally, the processor 270 may include one or more processing units; preferably, the processor 270 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface and application program, and the modem processor mainly processes wireless communication. It is understandable that the above-mentioned modem processor may not be integrated into the processor 270. In some embodiments, the processor and the memory may be implemented on a single chip, and in some embodiments, they may also be implemented separately on separate chips. The processor 270 may also be used to generate corresponding operation control signals, send them to corresponding components of the computing and processing device, read and process data in the software, especially read and process data and programs in the memory 220, so that each functional module therein performs corresponding functions, thereby controlling the corresponding components to act according to the requirements of the instructions.

[0068] Among them, the memory 220 can be used to store software codes related to the automated process generation method, the processor 270 can execute the steps of the automated process generation method, and can also schedule other units (such as the above-mentioned input unit 230 and display unit 240) to implement corresponding functions.

[0069] The radio frequency unit 210 (optional) can be used for receiving and sending information or receiving and sending signals during a call, for example, after receiving the downlink information of the base station, it is sent to the processor 270 for processing; in addition, the designed uplink data is sent to the base station. Generally, the radio frequency unit 210 includes but is not limited to an antenna, at least one amplifier, a transceiver, a coupler, a low noise amplifier (LowNoiseAmplifier, LNA), a duplexer, etc. In addition, the radio frequency unit 210 can also communicate with network devices and other devices through wireless communication. The wireless communication can use any communication standard or protocol, including but not limited to Global System of Mobile communication (Global System of Mobile communication, GSM), General Packet Radio Service (General PacketRadio Service, GPRS), Code Division Multiple Access (Code Division Multiple Access, CDMA), Wideband Code Division Multiple Access (Wideband Code Division Multiple Access, WCDMA), Long Term Evolution (Long Term Evolution, LTE), email, Short Messaging Service (SMS), etc.

[0070] In the embodiment of the present application, the radio frequency unit 210 can send data to other devices, and can also receive data sent by other devices. It should be understood that the radio frequency unit 210 is optional, and can be replaced by other communication interfaces, such as a network port.

[0071] The terminal 101 also includes a power supply 290 (such as a battery) for supplying power to various components. Preferably, the power supply can be logically connected to the processor 270 through a power management system, so that the power management system can manage functions such as charging, discharging, and power consumption.

[0072] The terminal 101 also includes an external interface 280 , which may be a standard Micro USB interface or a multi-pin connector, and may be used to connect the terminal 101 to communicate with other devices, or to connect a charger to charge the terminal 101 .

[0073] Although not shown, the terminal 101 may also include a flashlight, a wireless fidelity (WiFi) module, a Bluetooth module, sensors with different functions, etc., which will not be described in detail here.

[0074] Next, the product form of the above server is described.

[0075] Figure 3A structural diagram of the above server is provided, such as Figure 3 As shown, the server may include a bus 301, a processor 302, a communication interface 303, and a memory 304. The processor 302, the memory 304, and the communication interface 303 communicate with each other via the bus 301.

[0076] The bus 301 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0077] The processor 302 may be any one or more of a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor (MP), or a digital signal processor (DSP).

[0078] The memory 304 may include a volatile memory, such as a random access memory (RAM). The memory 304 may also include a non-volatile memory, such as a read-only memory (ROM), a flash memory, a hard disk drive (HDD), or a solid state drive (SSD).

[0079] The memory 304 may be used to store software codes related to the automated process generation method. The processor 302 may call the software codes stored in the memory 304 or schedule other units to implement corresponding functions.

[0080] The processors in the above-mentioned terminal 101 and server 102 (for example, processor 270 and processor 302) can be hardware circuits (such as application specific integrated circuit (ASIC), field-programmable gate array (FPGA), general-purpose processor, digital signal processor (DSP), microprocessor or microcontroller, etc.), or a combination of these hardware circuits. For example, the processor can be a hardware system with the function of executing instructions, such as CPU, DSP, etc., or a hardware system without the function of executing instructions, such as ASIC, FPGA, etc., or a combination of the above-mentioned hardware systems without the function of executing instructions and hardware systems with the function of executing instructions.

[0081] Next, the automated process generation method provided by the present application is introduced through the following embodiments.

[0082] See also Figure 4 , shows a schematic diagram of a process flow of an automated process generation method provided in an embodiment of the present application, the method may include:

[0083] Step S401: monitoring the natural language input by the user.

[0084] Step S402: when the natural language describing the target task input by the user is monitored, the natural language describing the target task is summarized and sorted to obtain key information of the target task.

[0085] Users can create a new session and directly use natural language to describe the target task, which is the user's operational requirement.

[0086] Exemplarily, the natural language input by the user to describe the target task is “Please help me automate the daily sales data, import the data from the spreadsheet into the database, and generate a daily sales report”.

[0087] Considering that the natural language input by the user contains a lot of invalid information (such as redundant details, lengthy explanations, colloquial expressions, irrelevant content, etc.), in order to avoid the impact of these invalid information on the generation of automated processes, when monitoring the natural language input by the user to describe the target task, this application summarizes and organizes the natural language describing the target task, filters out the invalid information in the user description, and extracts the core requirements, thereby obtaining the key information of the target task.

[0088] The target mission key information may include mission content, mission objectives, and operation steps. In a possible implementation, the target mission key information may be a structured natural language, specifically in the following form:

[0089] Task (summary, step [opt1, opt2, opt3, …])

[0090] Among them, Task represents the target task, summary represents the summary of the main line of the task, which briefly summarizes the content and purpose of the task, step represents a set of operation steps, each operation step in the operation step set includes several options (opt1, opt2, opt3, ...), and the several options contained in any operation step describe the specific operation of the operation step in detail.

[0091] In one possible implementation, a fine-tuned finite-scale language model can be used to summarize and organize the natural language describing the target task to obtain key information of the target task.

[0092] The fine-tuned finite-scale language model is obtained by fine-tuning a pre-trained finite-scale language model using training data corresponding to a variety of automation tasks. The training data corresponding to an automation task includes a natural language describing the automation task and corresponding task key information.

[0093] Step S403: According to the automation process component knowledge base, components matching the operation steps in the target task key information are screened from the automation process component library to obtain a matching component set and matching component description information.

[0094] The automation process component knowledge base contains description information of each component in the automation process component library.

[0095] The description information of any component included in the automation process component library includes the name of the component and the component function description of the component, wherein the component function description of any component includes: the specific function of the component, the input parameters of the component, the output parameters of the component and the usage scenario of the component.

[0096] This application pre-standardizes the annotations of each component in the automation process component library, that is, for each component in the automation process component library, a relatively standardized component name and component function are annotated. The component name should be concise and clear, and can reflect the core function of the component. The component function is described from four aspects, namely the specific function of the component (the specific function that the component can achieve), the input parameters of the component (the input parameters required by the component), the output parameters of the component (the output parameters generated by the component) and the usage scenario of the component (the typical usage scenario of the component).

[0097] The description information of any component included in the automation process component library can be a structured natural language, and the specific form is as follows:

[0098] component (name, description (function, input, output, example) )

[0099] Among them, component represents the component, name represents the component name, description represents the component function description, function represents the specific function of the component, input represents the input parameter of the component, output represents the output parameter of the component, and example represents the typical usage scenario of the component.

[0100] The description information of each component contained in the automation process component library constitutes the automation process component knowledge base. This application performs component matching based on the automation process component knowledge base. Since the automation process component knowledge base contains the standardized component description of each component in the automation process component library, therefore, performing component matching based on the automation process component knowledge base can achieve efficient and accurate matching of components.

[0101] When a matching component set is obtained through component matching, matching component description information can be obtained. The matching component description information includes description information of each component included in the matching component set.

[0102] In a possible implementation, the matching component description information may be in a structured natural language, and the specific form is as follows:

[0103] Components [compnent1 (name1, description1 (function1, input1,output1, example1) ), component2 (name2, description2 (function2, input2,output2, example2) ), component3(name3, description3 (function3, input3,output3, example3) ), …]

[0104] Among them, compnent1 (name1, description1 (function1, input1, output1,example1) ) represents the description information of component 1 in the matching component set, component2 (name2, description2(function2, input2, output2, example2) ) represents the description information of component 2 in the matching component set, component3(name3, description3 (function3, input3, output3, example3) ) represents the description information of component 3 in the matching component set, and so on.

[0105] Step S404: using the large language model, based on the target task key information and matching component description information, obtaining automated process generation guidance information.

[0106] It should be noted that this application will summarize and organize the natural language describing the target task to obtain the key information of the target task and provide it to the large language model (rather than directly providing the natural language describing the target task to the large language model), so that the large language model can more accurately understand the user's operational intentions, thereby improving the generation accuracy and efficiency of the automated process.

[0107] Specifically, using a large language model, based on target task key information and matching component description information, the process of obtaining automated process generation guidance information may include:

[0108] Step S4041: organize the target task key information, matching component description information and preset instruction prompt information into prompt instructions.

[0109] The target task key information (the aforementioned Task) and the matching component description information (the aforementioned Components) may be organized into a structured natural language (Task, Components), and then the structured natural language (Task, Components) and the preset instruction prompt information may be organized into prompt instructions.

[0110] Among them, the prompt instruction is used to prompt the large language model to generate guidance information for the automated process based on the key information of the target task and the matching component description information.

[0111] Step S4041: input the prompt instruction into the large language model to obtain the automated process generation guidance information output by the large language model.

[0112] After the prompt instruction is input into the large language model, the large language model first infers the user's operation intention based on the key information of the target task (the above-mentioned Task) and the matching component description information (the above-mentioned Components) in the prompt instruction, and then generates an automated process based on the inferred user operation intention and the key information of the target task (the above-mentioned Task) to generate guidance information and output it.

[0113] In one possible implementation, the automated process generation guidance information may be a structured natural language, and the automated process generation guidance information may include user operation intentions and component assembly information. Of course, this embodiment is not limited to this. For example, the automated process generation guidance information may also only include component assembly information.

[0114] The component assembly information is used to indicate how to assemble the components in the matching component set.

[0115] In a possible implementation, the component assembly information may be structured code data, and the structured code data may be code data of one of the following three execution modes, or may be code data of a combination of multiple execution modes of the following three execution modes.

[0116] The first execution mode: sequential execution

[0117] {

[0118] "steps": [

[0119] {

[0120] "step":componentName1,

[0121] “operation”: operation1

[0122] “inputParams”: { inputParam1, inputParam2, …}

[0123] “outputParam”: {outputParam1, outputParam2, …}

[0124] } ,

[0125] {

[0126] "step":componentName2,

[0127] “operation”: operation2

[0128] “inputParams”: { inputParam3, inputParam3, …}

[0129] "outputParam": {outputParam4, outputParam4, …}

[0130] }, … ]

[0132] }

[0133] Among them, the object array of "steps" is the execution order, the objects are components, "step" is the component name of the component to be used, "operation" is the operation of the component, "inputParams" is the input parameter of the component, and "outputParam" is the output parameter of the component.

[0134] The second execution method: loop execution

[0135] {

[0136] "step": LoopComponent,

[0137] "operation": "loop",

[0138] "inputParams": { "iterations": iterations ,"condition": condition}

[0139] },

[0140] "steps": [

[0141] {

[0142] "step": componentName1,

[0143] "operation": operantionName1,

[0144] “inputParams”: { inputParam1, inputParam2, …}

[0145] “outputParam”: {outputParam1, outputParam2, …}

[0146] } ]

[0148] }

[0149] Among them, "step": LoopComponent is the loop operation component, "operation" is the corresponding operation, in "inputParams", "iterations" indicates the number of loops, "condition" indicates the condition for maintaining the loop, and the "steps" below are the process automation components in the loop.

[0150] The third execution method: conditional execution

[0151] {

[0152] "step": ConditionComponent,

[0153] "operation": stepOperation,

[0154] "inputParams": { inputParam1, inputParam2, …},

[0155] "outputParam": {outputParam1, outputParam2, …},

[0156] "branches": [

[0157] {

[0158] "condition": Condition1,

[0159] "steps": [

[0160] {

[0161] "step": component1,

[0162] "operation": operation1,

[0163] "inputParams": { inputParam1, inputParam2, …},

[0164] “outputParam”: {outputParam1, outputParam2, …}

[0165] } ]

[0167] },

[0168] {

[0169] "condition": condition2,

[0170] "steps": [

[0171] {

[0172] "step": component2,

[0173] "operation": "generate_report",

[0174] "inputParams": { inputParam1, inputParam2, …},

[0175] “outputParam”: {outputParam1, outputParam2, …}

[0176] } ]

[0178] } ]

[0180] }

[0181] Among them, "step": ConditionComponent in the first line is the name of the conditional component, "operation" in the second line is the operation description, "inputParams" in the third line and "outputParam" in the fourth line are the input parameters and output parameters of the conditional component respectively, and the "branches" below are the branches of the conditional component. "condition": Condition1 means that in the case of Condition1, the component operations in the "steps" below are called sequentially, and "condition": condition2 means that in the case of Condition2, the component operations in the "steps" below are called sequentially.

[0182] Step S405: assembling each component in the matching component set according to the component assembly information to obtain an automated process.

[0183] Specifically, the process of assembling each component in the matching component set according to the component assembly information to obtain an automated process may include: determining the assembly order of each component in the matching component set and the parameter configuration of each component by parsing the component assembly information (such as the structured code data mentioned above), and assembling each component in the matching component set according to the assembly order of each component in the matching component set and the parameter configuration of each component, and combining the description information of each component in the matching component set to obtain an automated process.

[0184] The automated process generation method provided by the embodiment of the present application can monitor the natural language input by the user. When the natural language describing the target task input by the user is monitored, the natural language describing the target task is first summarized and sorted to obtain the key information of the target task, and then the components matching the operation steps contained in the key information of the target task are screened from the automated process component library according to the automated process component knowledge base to obtain a matching component set and obtain the matching component description information, and then the large language model is used to obtain the automated process generation guidance information based on the key information of the target task and the matching component description information, and finally, according to the component assembly information contained in the automated process generation guidance information, the components in the matching component set are assembled to obtain the automated process. The automated process generation method provided by the embodiment of the present application can automatically generate an automated process according to the natural language describing the target task input by the user, which greatly improves the generation efficiency of the automated process compared to the solution of manually arranging and designing the automated process. In addition, since the automated process generation method provided by the embodiment of the present application does not require manual arrangement and design, the influence of human subjective factors on the generation of the automated process is avoided.

[0185] In order to obtain a higher quality automated process, this application embodiment provides another automated process generation method, see Figure 5 , showing a flow chart of the automated process generation method, which may include:

[0186] Step S501: monitoring the natural language input by the user.

[0187] Step S502: when the natural language describing the target task input by the user is monitored, the natural language describing the target task is summarized and sorted to obtain key information of the target task.

[0188] Among them, the key information of the target task includes task content, task objectives and operation steps.

[0189] Step S503: According to the automation process component knowledge base, components matching the operation steps included in the target task key information are screened from the automation process component library to obtain a matching component set and matching component description information.

[0190] The automated process component knowledge base includes description information of each component in the automated process component library, and the matching component description information includes description information of each component in the matching component set.

[0191] Step S504: using the large language model, based on the target task key information and matching component description information, obtaining automated process generation guidance information.

[0192] Among them, the target task key information and matching component description information can both be structured natural language.

[0193] Among them, the automated process generates guidance information including component assembly information, and may also include the user operation intention inferred by the large language model based on the target task key information and the matching component description information. The component assembly information is used to indicate how to assemble the components in the matching component set.

[0194] Step S505: assembling each component in the matching component set according to the component assembly information to obtain an automated process.

[0195] The specific implementation process and related description of steps S501 to S505 can refer to the specific implementation process and related description of steps S401 to S405 in the above embodiment, which will not be repeated in this embodiment.

[0196] Step S506: Output the automation process.

[0197] After obtaining the automated process, the automated process can be exported for users to preview and test.

[0198] In a possible implementation, for the outputted automated process, it is possible to check whether there are any components with missing parameters in the automated process. If there are any components with missing parameters in the automated process, a natural language describing the missing parameters is outputted so that the user can supplement the parameters, wherein the natural language describing the missing parameters is used to indicate the components with missing parameters, the parameters missing from the components with missing parameters, and the impact of the missing parameters on the automated process. The above process not only enables users to accurately understand the deficiencies in the automated process, but also provides clear guidance to help users quickly and effectively improve the automated process. This interactive method improves the user experience, makes the entire automated process configuration process more intuitive and efficient, and ensures the integrity and functionality of the automated process finally generated.

[0199] Step S507a: if an adjustment suggestion for the automation process input by the user is monitored, the automation process generation guidance information is re-acquired based on the adjustment suggestion and the automation process generation guidance information acquired last time using the large language model.

[0200] After outputting the automation process, the present application can continuously monitor the natural language input by the user. When the user inputs an adjustment suggestion for the automation process, if the adjustment suggestion does not include a suggestion for a missing component, the large language model is used to re-acquire the automation process generation guidance information based on the adjustment suggestion and the automation process generation guidance information obtained last time. Specifically, the adjustment suggestion and the automation process generation guidance information obtained last time are first organized into a structured natural language (Task (userOperationIntent, Assembly information for components), userFeedback), and then the preset instruction prompt information and the structured natural language (Task (userOperationIntent, Assembly information for components), userFeedback) are organized into a prompt instruction (the prompt instruction is used to prompt the large language model to re-generate the automation process generation guidance information based on the adjustment suggestion and the automation process generation guidance information obtained last time), and the prompt instruction is input into the large language model to obtain the automation process generation guidance information output by the large language model. Among them, Task (userOperationIntent, Assemblyinformation for components) represents the last obtained automated process generation guidance information, userOperationIntent represents the user's operation intention, Assembly information for components represents component assembly information (structured code data), and userFeedback represents the user's adjustment suggestions.

[0201] When the user inputs an adjustment suggestion for the automated process, if the adjustment suggestion includes a suggestion for a missing component, first determine the possible missing component from the automated process component library based on the adjustment suggestion, and obtain the description information of the possible missing component. Then, use the large language model to re-acquire the automated process generation guidance information based on the modification suggestion, the description information of the possible missing component, and the automated process generation guidance information obtained last time.

[0202] In a possible implementation, the fine-tuned finite-scale language model can be used to analyze the adjustment suggestions, and then determine the possible missing components from the automation process component library based on the automation process component knowledge base. After determining the possible missing components, the adjustment suggestions, the description information of the possible missing components, and the automation process generation guidance information obtained last time can be organized into a structured natural language, and then the structured natural language and the preset instruction prompt information are organized into prompt instructions and input into the large language model to obtain the automation process generation guidance information output by the large language model.

[0203] Step S507b: If the natural language input by the user indicating that no adjustment is required is monitored, the generation process of the automation process is terminated.

[0204] Step S508a: assembling components according to the component assembly information included in the re-acquired automated process generation guidance information to obtain an automated process.

[0205] After assembling the components involved in the assembly information according to the assembly information of the components included in the re-acquired automated process generation guidance information to obtain the automated process, step S506 and subsequent steps may be executed, that is, after assembling the components according to the assembly information of the components included in the re-acquired automated process generation guidance information to obtain the automated process, waiting for the user to provide feedback on the automated process, if natural language inputted by the user indicating that no adjustment is required is monitored, the generation process of the automated process is terminated, if adjustment suggestions inputted by the user are monitored, the large language model is continued to be used to re-acquire the automated process generation guidance information based on the adjustment suggestions (all adjustment suggestions, including the adjustment suggestions monitored this time and the historical adjustment suggestions) and the automated process generation guidance information obtained last time, and then, assembling the components according to the assembly information of the components included in the latest automated process generation guidance information to obtain the latest automated process, after obtaining the latest automated process, the latest automated process may be output, and the user's feedback on the latest automated process is continuously monitored, ..., until the user's natural language inputted by the user indicating that no adjustment is required is monitored.

[0206] The automated process generation method provided by the present application can automatically generate an automated process based on the natural language input by the user to describe the target task. After the automated process is generated for the first time, the automated process can be continuously adjusted and optimized by interacting with the user, thereby ensuring that the automated process finally generated can meet the user's needs. For the user, only the operation requirements and modification suggestions are proposed through natural language to generate the required automated process. The automated process generation method provided by the present application improves the generation efficiency of the automated process, enhances the flexibility and adaptability of the automated process generation, and can create and optimize the automated process more easily, quickly and intuitively for the user.

[0207] The above introduces the automated process generation method provided in the embodiment of the present application. The following will introduce the device corresponding to the automated process generation method.

[0208] See also Figure 6 , Figure 6 A structural diagram of an intelligent agent provided in an embodiment of the present application, the intelligent agent may include: an input monitoring module 601, a task description processing module 602, a component matching module 603, an automated process generation guidance information acquisition module 604 and a component assembly module 605.

[0209] The input monitoring module 601 is used to monitor the natural language input by the user.

[0210] The task description processing module 602 is used to summarize and organize the natural language describing the target task when the input monitoring module 601 monitors the natural language describing the target task input by the user to obtain key information of the target task.

[0211] Among them, the key information of the target task includes task content, task objectives and operation steps.

[0212] The component matching module 603 is used to filter components matching the operation steps from the automation process component library according to the automation process component knowledge base to obtain a matching component set and obtain matching component description information.

[0213] The automation process component knowledge base includes description information of each component in the automation process component library, and the matching component description information includes description information of each component in the matching component set.

[0214] The automatic process generation guidance information acquisition module 604 is used to use the large language model to acquire the automatic process generation guidance information based on the target task key information and the matching component description information.

[0215] The automated process generation guidance information includes component assembly information, and the component assembly information is used to indicate how to assemble the components in the matching component set.

[0216] The component assembly module 605 is used to assemble the components in the matching component set according to the component assembly information to obtain an automated process.

[0217] In a possible implementation, when the task description processing module 602 summarizes and organizes the natural language describing the target task to obtain key information of the target task, it is specifically used to:

[0218] The natural language describing the target task is summarized and organized using a fine-tuned finite-scale language model to obtain the key information of the target task.

[0219] The fine-tuned finite-scale language model is obtained by fine-tuning a pre-trained finite-scale language model using training data corresponding to a variety of automation tasks. The training data corresponding to an automation task includes a natural language describing the automation task and corresponding task key information.

[0220] In a possible implementation, the target task key information and the matching component description information are both structured natural languages.

[0221] The automatic process generation guidance information acquisition module 604 is specifically used to:

[0222] Organizing the target task key information, the matching component description information and the preset instruction prompt information into prompt instructions, wherein the prompt instructions are used to prompt the large language model to generate guidance information for the automated process according to the target task key information and the matching component description information;

[0223] The prompt instruction is input into the large language model to obtain the automatic process generation guidance information output by the large language model.

[0224] In a possible implementation, the component assembly information is structured code data.

[0225] The component assembly module 605 assembles the components in the matching component set according to the component assembly information to obtain an automated process, specifically for:

[0226] By parsing the structured code data, the assembly order of each component in the matching component set and the parameter configuration of each component are determined;

[0227] According to the assembly order of each component in the matching component set and the parameter configuration of each component, and in combination with the matching component description information, each component in the matching component set is assembled to obtain an automated process.

[0228] In one possible implementation, the description information of any component contained in the automation process component library includes: the component name of the component and the component function description of the component, wherein the component function description of the component includes: the specific function of the component, the input parameters of the component, the output parameters of the component and the usage scenario of the component.

[0229] In a possible implementation, the automated process generates guidance information further including: the user operation intention inferred by the large language model based on the target task key information and the matching component description information.

[0230] The intelligent agent provided in the embodiment of the present application may also include: an automated process output module.

[0231] The automation process output module is used to output the automation process.

[0232] The automated process generation guidance information acquisition module 604 is also used to use the large language model to re-acquire the automated process generation guidance information based on the adjustment suggestions and the automated process generation guidance information acquired last time when the input monitoring module 601 monitors the user's input of adjustment suggestions for the automated process.

[0233] The component assembly module 605 is further used to assemble components according to the component assembly information included in the automated process generation guidance information re-acquired by the automated process generation guidance information acquisition module 604, so as to obtain a new automated process.

[0234] In a possible implementation, the intelligent agent provided by the embodiment of the present application may also include: an automated process inspection module.

[0235] The automated process check module is used to check whether there are any components with missing parameters in the automated process output by the automated process output module. If there are any components with missing parameters in the automated process output by the automated process output module, the module outputs a natural language describing the missing parameters so that the user can supplement the parameters. The natural language describing the missing parameters is used to indicate the components with missing parameters, the missing parameters of the components with missing parameters, and the impact of the missing parameters on the automated process.

[0236] In a possible implementation, the intelligent agent provided by the embodiment of the present application may also include: a missing component determination module.

[0237] The missing component determination module is used to determine the possible missing components from the automation process component library according to the adjustment suggestion when the adjustment suggestion includes a suggestion for missing components, and obtain description information of the possible missing components.

[0238] The automatic process generation guidance information acquisition module 604, when reacquiring the automatic process generation guidance information based on the adjustment suggestion and the automatic process generation guidance information acquired last time using the large language model, is specifically used to:

[0239] The large language model is used to retrieve the automated process generation guidance information based on adjustment suggestions, description information of possible missing components, and the automated process generation guidance information obtained last time.

[0240] The intelligent agent provided in the embodiment of the present application can automatically generate an automated process based on the natural language describing the target task input by the user. Compared with the solution of manually arranging and designing the automated process, the generation efficiency of the automated process is greatly improved. In addition, since the automated process generation method provided in the embodiment of the present application does not require manual arrangement and design, the influence of artificial subjective factors on the generation of the automated process is avoided. After the first generation of the automated process, the intelligent agent provided in the embodiment of the present application can continuously adjust and optimize the automated process by interacting with the user, thereby ensuring that the ultimately generated automated process can meet the needs of the user. The intelligent agent provided in the embodiment of the present application improves the generation efficiency of the automated process, enhances the flexibility and adaptability of the automated process generation, and can create and optimize the automated process more easily, quickly and intuitively for users.

[0241] The present application also provides an automated process generation system. Figure 7 , showing a schematic diagram of the structure of the automated process generation system, the automated process generation system may include: an intelligent agent 701 and a large language model module 702.

[0242] Intelligent agent 701 is used to monitor the natural language input by the user. When monitoring the natural language input by the user to describe the target task, it summarizes and organizes the natural language describing the target task to obtain key information of the target task. The key information of the target task includes task content, task objectives and operation steps. According to the automated process component knowledge base, components matching the operation steps are screened from the automated process component library to obtain a set of matching components and obtain matching component description information.

[0243] Among them, the automated process component knowledge base contains the description information of each component in the automated process component library. The description information of any component includes the component name and the component function description of the component. The component function description of the component includes the specific function of the component, the input parameters of the component, the output parameters of the component and the usage scenario of the component.

[0244] The matching component description information includes description information of each component in the matching component set.

[0245] The large language model module 702 is used to generate automatic process generation guidance information based on the target task key information and matching component description information.

[0246] The automated process generation guidance information includes component assembly information, and the component assembly information is used to indicate how to assemble components in the matching component set;

[0247] The intelligent agent 701 is also used to assemble the components in the matching component set according to the component assembly information to obtain an automated process.

[0248] In one possible implementation, agent 701 is integrated with a fine-tuned finite-scale language model.

[0249] When the agent 701 summarizes and organizes the natural language describing the target task to obtain the key information of the target task, it is specifically used to:

[0250] By using its own integrated fine-tuned finite-scale language model, the natural language describing the target task is summarized and organized to obtain the key information of the target task.

[0251] The fine-tuned finite-scale language model is obtained by fine-tuning a pre-trained finite-scale language model using training data corresponding to a variety of automation tasks. The training data corresponding to an automation task includes a natural language describing the automation task and corresponding task key information.

[0252] In a possible implementation, the target mission key information and the matching component description information are both structured natural languages, and the component assembly information included in the automated process generation guidance information is structured code data.

[0253] When the agent 701 assembles each component in the matching component set according to the component assembly information to obtain an automated process, it is specifically used to:

[0254] By parsing structured code data, the assembly order of each component in the matching component set and the parameter configuration of each component are determined; according to the assembly order of each component in the matching component set and the parameter configuration of each component, combined with the matching component description information, each component in the matching component set is assembled to obtain an automated process.

[0255] Considering that the large language model uses natural language for input and output, it cannot directly generate automated processes (although the large language model can understand and generate text for dialogue with humans, it cannot directly generate automated processes. Automated processes usually require the use of specialized programming languages ​​or tools to define specific operating steps and logic. For example, assembling automated processes in software systems is beyond the capabilities of large language models because they not only need to understand the semantics of the task, but also need to have the ability to execute complex instructions and operations). In addition, the natural language input by the user contains a lot of invalid information. Directly inputting the natural language input by the user into the large language model will affect the judgment of the large language model. In view of this, the present application uses the intelligent agent 701 as the key intermediate layer between the user and the large language model module 702. It extracts the key information of the target task from the natural language describing the target task input by the user through an integrated finite-scale language model, and screens the components matching the operation steps from the automated process component library to obtain the matching component description information, and provides the key information of the target task and the matching component description information to the large language model module 702. After obtaining the automated process generation guidance information returned by the large language model module 702, the intelligent agent 701 assembles the matching components according to the component assembly information in the automated process generation guidance information, thereby obtaining the automated process.

[0256] In a possible implementation, the automated process generates guidance information further including: the large language model module 702 infers the user operation intention based on the target task key information and the matching component description information.

[0257] The intelligent agent 701 is also used to output the automation process, and when monitoring the adjustment suggestions for the automation process input by the user, it provides the adjustment suggestions and the automation process generation guidance information obtained last time to the large language model module 702.

[0258] The large language model module 702 is also used to re-acquire the automated process generation guidance information based on the adjustment suggestion and the automated process generation guidance information acquired last time.

[0259] The intelligent agent 701 is also used to assemble components according to the component assembly information contained in the automated process generation guidance information re-acquired by the large language model module 702, so as to obtain a new automated process.

[0260] In a possible implementation, the agent 701 is also used to check whether there are any components in the output automation process that are missing parameters. If there are any components in the output automation process that are missing parameters, it outputs natural language describing the missing parameters so that the user can supplement the parameters.

[0261] The natural language describing the missing parameters is used to indicate the component with the missing parameters, the missing parameters of the component with the missing parameters, and the impact of the missing parameters on the automation process.

[0262] In one possible implementation, the intelligent agent 701 is also used to determine the possible missing components from the automated process component library based on the adjustment suggestions when the adjustment suggestions include suggestions for missing components, and obtain description information of the possible missing components, and provide the description information of the possible missing components to the large language model module 702.

[0263] When reacquiring the automated process generation guidance information based on the adjustment suggestion and the automated process generation guidance information acquired last time, the large language model module 702 is specifically used to:

[0264] Re-acquire the automated process generation guidance information based on the adjustment suggestions, description information of the components that may be missing, and the automated process generation guidance information that was obtained last time.

[0265] The automated process generation system provided by the embodiment of the present application, the intelligent agent can monitor the natural language input by the user, and when the natural language describing the target task is monitored by the user, the automated process can be automatically generated with the assistance of the large language model module, that is, the user only needs to describe his needs through natural language, and the intelligent agent can automatically match the components, and complete the assembly of the matching components with the assistance of the large language model module to obtain the automated process. After the automated process is generated for the first time, the intelligent agent can continuously adjust and optimize the automated process with the assistance of the large language model module by interacting with the user, thereby ensuring that the ultimately generated automated process can meet the user's needs. The automated process generation system provided by the embodiment of the present application improves the generation efficiency of the automated process, enhances the flexibility and adaptability of the automated process generation, and can create and optimize the automated process more easily, quickly and intuitively for users.

[0266] An embodiment of the present application also provides an electronic device, which may include: at least one processor, at least one communication interface, at least one memory and at least one communication bus.

[0267] In the embodiment of the present application, the number of the processor, the communication interface, the memory, and the communication bus is at least one, and the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0268] The processor may be a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application;

[0269] The memory may include a high-speed RAM memory, and may also include a non-volatile memory, such as at least one disk memory;

[0270] The memory stores a program, and the processor can call the program stored in the memory, and the program is used to implement the steps of the automatic process generation method provided in the above embodiment.

[0271] An embodiment of the present application also provides a computer storage medium, which carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement the steps of the automated process generation method provided in the above embodiment.

[0272] An embodiment of the present application also provides a computer program product, including computer-readable instructions. When the computer-readable instructions are executed on an electronic device, the electronic device implements the steps of the automated process generation method provided in the above embodiment.

[0273] It should also be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed over multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the drawings of the device embodiments provided by the present application, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines.

[0274] Through the description of the above implementation mode, the technicians in the field can clearly understand that the present application can be implemented by means of software plus necessary general hardware, and of course, it can also be implemented by special hardware including special integrated circuits, special CPUs, special memories, special components, etc. In general, all functions completed by computer programs can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be various, such as analog circuits, digital circuits or special circuits. However, for the present application, software program implementation is a better implementation mode in more cases. Based on such an understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a readable storage medium, such as a computer floppy disk, a U disk, a mobile hard disk, a ROM, a RAM, a disk or an optical disk, etc., including a number of instructions to enable a computer device (which can be a personal computer, a training device, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0275] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product.

[0276] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from a website site, a computer, a training device, or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, training device, or data center. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a training device, a data center, etc. that includes one or more available media integrations. The available medium may be a magnetic medium, (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)), etc.

Claims

1. An automated process generation method, characterized in that: include: Monitor the natural language input by the user; When natural language describing the target task input by the user is monitored, the natural language describing the target task is summarized and sorted to obtain key information of the target task, wherein the key information of the target task includes task content, task objectives and operation steps; According to the automation process component knowledge base, components matching the operation steps are screened from the automation process component library to obtain a matching component set, and matching component description information is obtained, wherein the automation process component knowledge base includes description information of each component in the automation process component library, and the matching component description information includes description information of each component in the matching component set; Using the large language model, based on the target task key information and the matching component description information, obtaining automated process generation guidance information, wherein the automated process generation guidance information includes component assembly information, the component assembly information is used to indicate how to assemble components in the matching component set, and the automated process generation guidance information also includes the user operation intention inferred by the large language model based on the target task key information and the matching component description information; Assembling each component in the matching component set according to the component assembly information to obtain an automated process; The automated process generation method further comprises: outputting the automated process; When an adjustment suggestion for the automation process input by the user is monitored, the automation process generation guidance information is reacquired based on the adjustment suggestion and the automation process generation guidance information acquired last time by using the large language model; Assembling components according to the component assembly information included in the re-acquired automated process generation guidance information to obtain a new automated process; The automated process generation method further comprises: If the adjustment suggestion includes a suggestion for a missing component, determining the possibly missing component from the automation process component library according to the adjustment suggestion, and obtaining description information of the possibly missing component; The using the large language model to reacquire the automated process generation guidance information based on the adjustment suggestion and the automated process generation guidance information acquired last time includes: The large language model is used to reacquire the automated process generation guidance information based on the adjustment suggestion, the description information of the possibly missing component, and the automated process generation guidance information acquired last time.

2. The automated process generation method according to claim 1, characterized in that: The natural language describing the target task is summarized and sorted to obtain key information of the target task, including: Using a fine-tuned finite-scale language model, the natural language describing the target task is summarized and organized to obtain key information of the target task; The fine-tuned finite-scale language model is obtained by fine-tuning a pre-trained finite-scale language model using training data corresponding to a plurality of automation tasks, wherein the training data corresponding to an automation task includes a natural language describing the automation task and corresponding key task information.

3. The automated process generation method according to claim 1, characterized in that: The target task key information and the matching component description information are both structured natural languages; The using of the large language model, based on the target task key information and the matching component description information, to obtain the automatic process generation guidance information includes: Organizing the target task key information, the matching component description information and the preset instruction prompt information into a prompt instruction, wherein the prompt instruction is used to prompt the large language model to generate an automated process to generate guidance information according to the target task key information and the matching component description information; The prompt instruction is input into a large language model to obtain automated process generation guidance information output by the large language model.

4. The automated process generation method according to claim 1, characterized in that: The component assembly information is structured code data; The step of assembling each component in the matching component set according to the component assembly information to obtain an automated process includes: Determining the assembly order of each component in the matching component set and the parameter configuration of each component by parsing the structured code data; According to the assembly order of each component in the matching component set and the parameter configuration of each component, and in combination with the matching component description information, each component in the matching component set is assembled to obtain an automated process.

5. The automated process generation method according to claim 1, characterized in that: The description information of any component included in the automation process component library includes: the component name of the component and the component function description of the component; The component function description of the component includes: the specific function of the component, the input parameters of the component, the output parameters of the component and the usage scenario of the component.

6. The automated process generation method according to claim 1, characterized in that: The automated process generation method further comprises: Check if there are any components missing parameters in the output automation process; If there are components with missing parameters in the output automation process, natural language describing the missing parameters is output so that the user can supplement the parameters, wherein the natural language describing the missing parameters is used to indicate the components with missing parameters, the parameters missing from the components with missing parameters, and the impact of the missing parameters on the automation process.

7. An intelligent agent, characterized in that: include: Input monitoring module, task description processing module, component matching module, automated process generation guidance information acquisition module and component assembly module; The input monitoring module is used to monitor the natural language input by the user; The task description processing module is used to summarize and organize the natural language describing the target task when the input monitoring module monitors the natural language describing the target task input by the user to obtain key information of the target task, wherein the key information of the target task includes task content, task objectives and operation steps; The component matching module is used to screen the components matching the operation steps from the automation process component library according to the automation process component knowledge base to obtain a matching component set and obtain matching component description information, wherein the automation process component knowledge base contains description information of each component in the automation process component library, and the matching component description information includes description information of each component in the matching component set; The automated process generation guidance information acquisition module is used to use the large language model to acquire the automated process generation guidance information based on the target task key information and the matching component description information, wherein the automated process generation guidance information includes component assembly information, and the component assembly information is used to indicate how to assemble the components in the matching component set, and the automated process generation guidance information also includes the user operation intention inferred by the large language model based on the target task key information and the matching component description information; The component assembly module is used to assemble each component in the matching component set according to the component assembly information to obtain an automated process; The intelligent agent also includes: an automated process output module; The automation process output module is used to output the automation process; The automated process generation guidance information acquisition module is further configured to, when the input monitoring module monitors the user input of the adjustment suggestion for the automated process, re-acquire the automated process generation guidance information based on the adjustment suggestion and the automated process generation guidance information acquired last time using the large language model; The component assembly module is further used to assemble components according to the component assembly information included in the automated process generation guidance information re-acquired by the automated process generation guidance information acquisition module, so as to obtain a new automated process; The agent also includes: a missing component determination module; The missing component determination module is used to determine the possible missing components from the automation process component library according to the adjustment suggestion when the adjustment suggestion includes a suggestion for missing components, and obtain description information of the possible missing components; The automated process generation guidance information acquisition module is specifically used to utilize the large language model to re-acquire the automated process generation guidance information based on the adjustment suggestions, the description information of the possibly missing components, and the automated process generation guidance information acquired last time, based on the large language model.

8. An automated process generation system, characterized in that: include: Agent and large language model modules; The intelligent agent is used to listen to the natural language input by the user, and when listening to the natural language input by the user describing the target task, the intelligent agent summarizes and organizes the natural language describing the target task to obtain key information of the target task, wherein the key information of the target task includes task content, task objectives and operation steps, and according to the automated process component knowledge base, the intelligent agent screens the components matching the operation steps from the automated process component library to obtain a matching component set, and obtains matching component description information, wherein the automated process component knowledge base contains description information of each component in the automated process component library, and the matching component description information includes description information of each component in the matching component set; The large language model module is used to generate automatic process generation guidance information based on the target task key information and the matching component description information, wherein the automatic process generation guidance information includes component assembly information, and the component assembly information is used to indicate how to assemble the components in the matching component set, and the automatic process generation guidance information also includes the user operation intention inferred by the large language model module based on the target task key information and the matching component description information; The intelligent agent is further used to assemble each component in the matching component set according to the component assembly information to obtain an automated process; The intelligent agent is further used to output the automation process, and when monitoring the adjustment suggestions for the automation process input by the user, provide the adjustment suggestions and the automation process generation guidance information obtained last time to the large language model module; The large language model module is further used to reacquire the automated process generation guidance information based on the adjustment suggestion and the automated process generation guidance information acquired last time; The intelligent agent is further used to assemble components according to the component assembly information included in the automated process generation guidance information reacquired by the large language model module to obtain a new automated process; The agent is further configured to, when the adjustment suggestion includes a suggestion of a missing component, determine the possibly missing component from the automated process component library according to the adjustment suggestion, obtain description information of the possibly missing component, and provide the description information of the possibly missing component to the large language model module; When the large language model module re-acquires the automated process generation guidance information based on the adjustment suggestions and the automated process generation guidance information acquired last time, it is specifically used to re-acquire the automated process generation guidance information based on the adjustment suggestions, the description information of the possibly missing components and the automated process generation guidance information acquired last time.

Images