Recommendation agent system, recommendation method and recommendation device

Abstract

The invention discloses a recommendation agent system, a recommendation method and a recommendation device, and relates to the technical field of agents, a sensing module obtains first dialogue data and / or multi-modal environment data, and a memory module obtains historical man-machine interaction data related to user intention and a target object to be recommended; the planning module retrieves a reference workflow matched with the intention of the user, generates an initial agent workflow according to the data acquired by the sensing module, the historical man-machine interaction data, the target object to be recommended and the reference workflow, and calls the execution module to execute the initial agent workflow; judging whether the initial agent workflow is updated or not based on second dialogue data and / or return data generated in the execution process, if yes, updating the initial agent workflow, and executing the updated agent workflow from the updating position to complete recommendation. According to the method, the recommendation tasks can be dynamically planned in combination with multiple types of data, and the planning accuracy and recommendation accuracy are improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent agent technology, and in particular to a recommendation intelligent agent system, a recommendation method and a recommendation device. Background Art

[0002] The intelligent recommendation system is an important part of digital marketing business. It can filter out the most suitable products from massive information based on user needs and interests, thereby improving user satisfaction and loyalty.

[0003] The process of traditional intelligent recommendation systems is to obtain historical interaction data between users and front-end applications, and use the recommendation system model to obtain recommendation results based on historical interaction data. Since historical interaction data can only reflect users' past product preferences, the recommendation results obtained based on historical recommendation interaction data may be inconsistent with users' current product preferences, resulting in inaccurate recommendation results. Summary of the invention

[0004] In view of the above problems, the present application provides a recommendation agent system, a recommendation method and a recommendation device to achieve the purpose of improving the accuracy of recommendation results. The specific scheme is as follows: The first aspect of the present application provides a recommendation agent system, comprising: a perception module, a memory module, a planning module and an execution module; The perception module is used to obtain first dialogue data and / or multimodal environment data collected during the human-computer interaction process, wherein the first dialogue data includes at least one round of dialogue data; The memory module is used to obtain historical human-computer interaction data related to user intention and the target object to be recommended, wherein the user intention reflects the recommended object that the user prefers; The planning module is used to retrieve the reference workflow matching the user intention, and generate an initial agent workflow based on the data acquired by the perception module, the historical human-computer interaction data, the target object to be recommended and the reference workflow, wherein the reference workflow refers to the orchestration workflow used when the object in the scene related to the user intention is successfully recommended, or the manually orchestrated workflow after matching the user intention and verifying the recommendation effect, and the initial agent workflow is a workflow planned and orchestrated by the recommendation big model using the reference workflow as a thinking chain used for reasoning of the recommendation big model; The planning module is also used to call the execution module to execute the initial intelligent agent workflow, and determine whether to update the initial intelligent agent workflow based on the second dialogue data and / or return data generated during the execution process. If so, the initial intelligent agent workflow is updated based on the first dialogue data, the second dialogue data and the reference workflow, and the execution module is called to execute the updated intelligent agent workflow from the update position to complete the recommendation, wherein the second dialogue data is at least one round of dialogue data with the user based on a step-by-step dialogue guidance strategy.

[0005] In a possible implementation, when the planning module generates the initial agent workflow according to the data acquired by the perception module, the historical human-computer interaction data, the target object to be recommended and the reference workflow, it is specifically used to: Fill the data acquired by the perception module, the historical human-computer interaction data, the proposed recommended target object and the reference workflow into the first prompt word prompt template of the recommendation large model to obtain a first prompt instruction prompt; Inputting the first prompt instruction prompt into the recommendation large model to obtain the initial agent workflow output by the recommendation large model; Among them, the training data used by the recommendation model in the training stage includes: training multimodal environment data and / or training dialogue data, training historical human-computer interaction data, training target objects to be recommended, training reference workflows, and corresponding training initial intelligent agent workflows.

[0006] In a possible implementation, the training data used by the recommendation big model in the training phase also includes: preconfigured custom workflows and workflow generation rules, wherein the workflow generation rules are used to instruct the recommendation big model to follow the execution logic and execution order in the custom workflow to generate the required workflow.

[0007] In a possible implementation, before the memory module acquires the historical human-computer interaction data related to the user intention and the target object to be recommended, the perception module is further configured to: The user intention is determined based on the data acquired by the perception module.

[0008] In a possible implementation, when the perception module determines the user intention according to the data acquired by the perception module, it is specifically used to: Filling the data acquired by the perception module into the second prompt word prompt template of the recommendation large model to obtain a second prompt instruction prompt; Inputting the second prompt instruction prompt into the recommendation model to obtain the user intention output by the recommendation model; The training data used by the recommendation large model in the training phase includes: training multimodal environment data and / or training conversation data, and corresponding training user intentions.

[0009] In a possible implementation, when acquiring the historical human-computer interaction data and the target object to be recommended related to the user intention, the memory module is specifically used to: According to the user intention, acquiring the historical human-computer interaction data from a first memory bank; According to the user intention, the proposed recommended target object is obtained from the second memory library.

[0010] In a possible implementation, the multimodal environment data includes one or more of image data, video data, and audio data.

[0011] In a possible implementation, the historical human-computer interaction data is a content summary of historical conversation data between the user and the recommendation assistant.

[0012] A second aspect of the present application provides a recommendation method, which is applied to a recommendation agent system, comprising: Acquire first dialogue data and / or multimodal environment data collected during the human-computer interaction process, wherein the first dialogue data includes at least one round of dialogue data; Acquire historical human-computer interaction data related to user intention and target objects to be recommended, wherein the user intention reflects the recommended objects that the user prefers; Retrieve a reference workflow that matches the user intention, and generate an initial agent workflow based on the first dialogue data and / or the multimodal environment data, the historical human-computer interaction data, the target object to be recommended, and the reference workflow, wherein the reference workflow refers to an orchestration workflow used when successfully recommending an object in a scene related to the user intention, or a manually orchestrated workflow that matches the user intention and verifies the recommendation effect, and the initial agent workflow is a workflow that uses the reference workflow as a thinking chain used for reasoning with the recommendation big model and is planned and orchestrated by the recommendation big model; Execute the initial agent workflow, determine whether to update the initial agent workflow based on second dialogue data and / or return data generated during the execution process, and if so, update the initial agent workflow based on the first dialogue data, the second dialogue data and the reference workflow, and execute the updated agent workflow from the update position to complete the recommendation, wherein the second dialogue data is at least one round of dialogue data with the user based on a step-by-step dialogue guidance strategy.

[0013] The third aspect of the present application provides a recommendation device, which is applied to a recommendation agent system, comprising: A data acquisition unit, configured to acquire first dialogue data and / or multimodal environment data collected during the human-computer interaction process, wherein the first dialogue data includes at least one round of dialogue data; A memory retrieval unit, used to obtain historical human-computer interaction data related to user intention and target objects to be recommended, wherein the user intention reflects the recommended objects that the user prefers; A workflow orchestration unit, configured to retrieve a reference workflow matching the user intent, and generate an initial agent workflow according to the first dialogue data and / or the multimodal environment data, the historical human-computer interaction data, the target object to be recommended, and the reference workflow, wherein the reference workflow refers to an orchestration workflow used when successfully recommending an object in a scene related to the user intent, or a manually orchestrated workflow that matches the user intent and verifies the recommendation effect, and the initial agent workflow is a workflow planned and orchestrated by the recommendation large model using the reference workflow as a thinking chain used for reasoning of the recommendation large model; A dynamic programming unit is used to execute the initial intelligent agent workflow, determine whether to update the initial intelligent agent workflow based on the second dialogue data and / or return data generated during the execution process, and if so, update the initial intelligent agent workflow based on the first dialogue data, the second dialogue data and the reference workflow, and execute the updated intelligent agent workflow from the updated position to complete the recommendation, wherein the second dialogue data is at least one round of dialogue data with the user based on a step-by-step dialogue guidance strategy.

[0014] With the above technical solution, the recommendation intelligent agent system provided by the present application, the perception module obtains the first dialogue data and / or multimodal environment data collected during the human-computer interaction process, determines the user intention based on the obtained data, the memory module obtains the historical human-computer interaction data and the target object to be recommended related to the user intention, the planning module retrieves the reference workflow matching the user intention, generates an initial intelligent agent workflow based on the data obtained by the perception module, the historical human-computer interaction data, the target object to be recommended and the reference workflow, calls the execution module to execute the initial intelligent agent workflow, determines whether to update the initial intelligent agent workflow based on the second dialogue data and / or return data generated during the execution process, and if so, updates the initial intelligent agent workflow based on the first dialogue data, the second dialogue data and the reference workflow, and calls the execution module to execute the updated intelligent agent workflow from the update position to complete the recommendation.

[0015] Since the first conversation data and the multimodal environment data are data generated when the user needs a recommendation, the first conversation data and the multimodal environment data can better reflect the user's current preferences, and the historical human-computer interaction data can reflect the user's past interests. Therefore, the recommendation agent system provided in the present application refers to the first conversation data, multimodal environment data and historical human-computer interaction data at the same time when making recommendations, and can more accurately determine the recommendation objects that the user prefers, thereby improving the accuracy of recommendations.

[0016] The proposed target object provides the recommendation direction for the recommendation process. The initial intelligent agent workflow is planned based on the proposed target object, and then product recommendations are made based on the initial intelligent agent workflow. This can guide users to establish interest in the proposed target object during the recommendation process, thereby increasing the probability of successful recommendation of the proposed target object. At the same time, the existence of the proposed target object makes active recommendation possible.

[0017] Furthermore, the historical human-computer interaction data in this application is the user's long-term historical memory data. Combining the historical human-computer interaction data with the target object to be recommended enables this application to plan a long-term recommendation strategy and increase the probability of successful recommendation.

[0018] Furthermore, the present application can draw on the successful recommendation experience of objects in scenes related to user intentions, and use the reference workflow used when successfully recommending the object as the thinking chain used by the recommendation large model reasoning, so that the recommendation large model can draw on the thinking mode of successful recommendation, improve the accuracy of the automatic planning of the recommendation large model, and make recommendations based on the more accurate initial intelligent agent workflow arranged by the recommendation large model, thereby improving the accuracy of the recommendation. Even if the orchestration workflow used to successfully recommend objects in scenes related to user intentions cannot be retrieved, the present application can still use the manually orchestrated workflow that matches the user intention and verifies the recommendation effect as the reference workflow, so that the initial intelligent agent workflow generated based on the reference workflow can be more effectively recommended, thereby improving the recommendation success rate.

[0019] In addition, regardless of whether the current recommendation task is complex or not, this application can retrieve the corresponding reference workflow, and generate the initial intelligent agent workflow after learning the recommendation thinking mode of the reference workflow, so that the initial intelligent agent workflow can provide a better recommendation strategy for the current recommendation task, improving the recommendation accuracy and success rate. Compared with the existing recommendation scheme that can only solve simple recommendation tasks, this application can provide an accurate initial intelligent agent workflow to solve complex recommendation tasks, and has a wider range of applications.

[0020] Furthermore, the recommendation agent system provided by the present application can adopt a step-by-step conversational guidance strategy to conduct one or more rounds of conversations with the user to obtain second conversation data during the process of making recommendations based on the automatically planned initial agent workflow. The addition of the second conversation data and the return data generated during the execution of the initial agent workflow enables the present application to better understand the user's intention, and then dynamically update the initial agent workflow based on the understanding of the user's intention, so that the updated agent workflow is more in line with the user's intention, thereby improving the accuracy of the recommendation. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the accompanying drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale.

[0022] Figure 1 A schematic diagram of a system architecture provided for this application; Figure 2 A schematic diagram of an optional hardware structure of the terminal 100 provided in this application; Figure 3 A schematic diagram of the structure of a server 200 provided in this application; Figure 4 A schematic diagram of the structure of a recommendation agent system provided in this application; Figure 5 A schematic diagram of the interaction between various modules when performing dynamic planning based on the recommendation large model provided in this application; Figure 6 A flowchart of a recommended method provided for this application; Figure 7 A schematic diagram of the structure of a recommended device provided for this application; Figure 8 A schematic diagram of the structure of an electronic device provided in this application. DETAILED DESCRIPTION

[0023] 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.

[0024] 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.

[0025] 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.

[0026] See also Figure 1 , Figure 1 A schematic diagram of a system architecture is shown. The system may include a terminal 100 and a server 200. The server 200 may include one or more servers ( Figure 1 In the example, a server is included, and the server 200 can provide the method provided in the embodiment of the present application for one or more terminals.

[0027] Among them, an application corresponding to the third-party system can be installed on the terminal 100. The above application and web page can provide an interface. The terminal 100 can receive relevant parameters entered by the user on the interface and send the above parameters to the server 200. The server 200 can obtain processing results based on the received parameters and return the processing results to the terminal 100.

[0028] It should be understood that in some optional implementations, the terminal 100 can also complete the action of obtaining the processing result based on the received parameters by itself without the cooperation of the server, and the embodiments of the present application are not limited to this.

[0029] Next describe Figure 1 The product form of the mid-terminal 100; The terminal 100 in the embodiment of the present application 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 embodiment of the present application does not impose any restrictions on this.

[0030] Figure 2 An optional hardware structure diagram of the terminal 100 is shown.

[0031] refer to Figure 2 As shown, the terminal 100 may include a radio frequency unit 110, a memory 120, an input unit 130, a display unit 140, a camera 150 (optional), an audio circuit 160 (optional), a speaker 161 (optional), a microphone 162 (optional), an earphone jack 163 (optional), a processor 170, an external interface 180, a power supply 190 and other components. Those skilled in the art will appreciate that Figure 2 These are merely examples of terminals or multi-function devices and do not constitute limitations on the terminals or multi-function devices, which may include more or fewer components than those shown in the figures, or combinations of certain components, or different components.

[0032] The input unit 130 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 portable multifunctional device. Specifically, the input unit 130 may include a touch screen 131 (optional) and / or other input devices 132. The touch screen 131 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 170, and can receive and execute the command sent by the processor 170; the touch signal at least includes the touch point coordinate information. The touch screen 131 can provide an input interface and an output interface between the terminal 100 and the user. In addition, the touch screen can be implemented using multiple types such as resistive, capacitive, infrared and surface acoustic wave. In addition to the touch screen 131, the input unit 130 can also include other input devices. Specifically, other input devices 132 may include, but are not limited to, one or more of a physical keyboard, function keys (such as a volume control key, a switch key, etc.), a trackball, a mouse, a joystick, and the like.

[0033] Among them, the input device 132 can receive input data and the like.

[0034] The display unit 140 may be used to display information input by a user or information provided to a user, various menus of the terminal 100, an interactive interface, file display, and / or playback of any multimedia file.

[0035] The memory 120 can be used to store instructions and data. The memory 120 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, instructions required for at least one function, or their subsets and extensions. It can also include a non-volatile random access memory; provide the processor 170 with hardware, software and data resources including management of computing and processing equipment, and support control software and applications. It is also used for the storage of multimedia files, and the storage of running programs and applications.

[0036] The processor 170 is the control center of the terminal 100. It uses various interfaces and lines to connect various parts of the entire terminal 100. By running or executing instructions stored in the memory 120 and calling data stored in the memory 120, it executes various functions of the terminal 100 and processes data, thereby controlling the terminal device as a whole. Optionally, the processor 170 may include one or more processing units; preferably, the processor 170 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 170. 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 170 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 120, so that each functional module therein performs corresponding functions, thereby controlling the corresponding components to act according to the requirements of the instructions.

[0037] The memory 120 may be used to store software codes related to the recommendation method, the processor 170 may execute the steps of the recommendation method, and may also schedule other units (such as the above-mentioned input unit 130 and display unit 140) to implement corresponding functions.

[0038] The radio frequency unit 110 (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 170 for processing; in addition, the designed uplink data is sent to the base station. Generally, the RF circuit includes but is not limited to an antenna, at least one amplifier, a transceiver, a coupler, a low noise amplifier (Low Noise Amplifier, LNA), a duplexer, etc. In addition, the radio frequency unit 110 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 Packet Radio 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.

[0039] In this embodiment of the present application, the RF unit 110 can send data to the server 200 and receive processing results sent by the server 200.

[0040] It should be understood that the radio frequency unit 110 is optional and can be replaced by other communication interfaces, such as a network port.

[0041] The terminal 100 also includes a power supply 190 (such as a battery) for supplying power to various components. Preferably, the power supply can be logically connected to the processor 170 through a power management system, so that the power management system can manage functions such as charging, discharging, and power consumption.

[0042] The terminal 100 further includes an external interface 180 , which may be a standard Micro USB interface or a multi-pin connector, and may be used to connect the terminal 100 to communicate with other devices, or to connect a charger to charge the terminal 100 .

[0043] Although not shown, the terminal 100 may also include a flashlight, a wireless fidelity (WiFi) module, a Bluetooth module, sensors with different functions, etc., which are not described in detail here. Some or all of the methods described below may be applied in the following embodiments. Figure 2 In the terminal 100 shown.

[0044] Next describe Figure 1 The product form of the server 200; Figure 3 A structural diagram of a server 200 is provided, such as Figure 3 As shown, the server 200 includes a bus 201, a processor 202, a communication interface 203, and a memory 204. The processor 202, the memory 204, and the communication interface 203 communicate with each other via the bus 201.

[0045] The bus 201 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.

[0046] The processor 202 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).

[0047] The memory 204 may include a volatile memory, such as a random access memory (RAM). The memory 204 may also include a non-volatile memory, such as a read-only memory (ROM), a flash memory, a mechanical hard drive (HDD), or a solid state drive (SSD).

[0048] The memory 204 may be used to store software codes related to the recommendation method, and the processor 202 may execute the steps of the chip recommendation method and may also schedule other units to implement corresponding functions.

[0049] It should be understood that the above-mentioned terminal 100 and server 200 can be centralized or distributed devices, and the processors in the above-mentioned terminal 100 and server 200 (such as processor 170 and processor 202) can be hardware circuits (such as application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), general-purpose processors, DSPs, microprocessors or microcontrollers, etc.), or a combination of these hardware circuits. For example, the processor can be a hardware system with an instruction execution function, such as a CPU, DSP, etc., or a hardware system without an instruction execution function, such as an ASIC, FPGA, etc., or a combination of the above-mentioned hardware systems without an instruction execution function and hardware systems with an instruction execution function.

[0050] See also Figure 4 , is a structural diagram of a recommendation agent system provided in this application, such as Figure 4 As shown, the recommendation agent system includes a perception module 51, a memory module 53, a planning module 52 and an execution module 54.

[0051] The perception module 51 is used to obtain first dialogue data and / or multimodal environment data collected during the human-computer interaction process, wherein the first dialogue data includes at least one round of dialogue data; A memory module 53 is used to obtain historical human-computer interaction data related to user intention and the target object to be recommended, wherein the user intention reflects the recommended object that the user prefers; The planning module 52 is used to retrieve the reference workflow matching the user intention, and generate the initial agent workflow according to the data obtained by the perception module, the historical human-computer interaction data, the target object to be recommended and the reference workflow, wherein the reference workflow refers to the orchestration workflow used when the object in the scene related to the user intention is successfully recommended, or the manually orchestrated workflow after matching the user intention and verifying the recommendation effect, and the initial agent workflow is the workflow planned and orchestrated by the recommendation big model, which uses the reference workflow as the thinking chain used for the reasoning of the recommendation big model; The planning module 52 is also used to call the execution module to execute the initial intelligent agent workflow, and determine whether to update the initial intelligent agent workflow based on the second dialogue data and / or return data generated during the execution process. If so, the initial intelligent agent workflow is updated based on the first dialogue data, the second dialogue data and the reference workflow, and the execution module is called to execute the updated intelligent agent workflow from the update position to complete the recommendation, wherein the second dialogue data is at least one round of dialogue data with the user based on a step-by-step dialogue guidance strategy.

[0052] The above process is introduced below.

[0053] The recommendation agent system provided in the present application has a recommendation assistant set up in it, and the user can have one or more rounds of conversations with the recommendation assistant, and the conversation data of the first or more rounds of conversations serves as the first conversation data of the present application.

[0054] The perception module 51 in the present application can also obtain multimodal environmental data, which is environmental data related to the user's current recommendation task. Optionally, the multimodal environmental data includes one or more of image data, video data, and audio data.

[0055] In actual application scenarios, multimodal environmental data can be environmental data input by the user or other environmental data.

[0056] The above-mentioned environmental data input by the user is, for example, data such as images and videos that the user views or clicks on on the page browsed in this recommendation task. The data viewed or clicked by the user is the environmental data input by the user. Since the environmental data viewed and clicked by the user are often the data that the user is more concerned about and interested in, making recommendations based on these data helps improve the accuracy of recommendations.

[0057] Other environmental data include, for example, automatically generating conversation-related pictures, videos, audio and other data based on the conversation content during the conversation between the user and the recommendation assistant.

[0058] Of course, the multimodal environment data may also be other types, which are not specifically limited in this application.

[0059] The memory module 53 can obtain relevant historical human-computer interaction data and target objects to be recommended based on user intention.

[0060] Optionally, the historical human-computer interaction data may be historical conversation data between the user and the recommendation assistant, or may be a content summary of the historical conversation data between the user and the recommendation assistant. Here, the content summary refers to useful information in the historical conversation data, such as user preferences, user historical events (such as where to travel, what movies to watch, what awards to win, etc.), etc.

[0061] The above-mentioned proposed target object refers to the target object that the recommendation agent system intends to recommend to the user. In actual application scenarios, the recommendation agent system provided by this application may recommend the proposed target object to the user, or may not recommend the proposed target object to the user. For example, in some recommendation scenarios, the recommendation agent system finds that the user is interested in the proposed target object during the interaction with the user through the recommendation assistant, and then the proposed target object may be recommended to the user; in other recommendation scenarios, the recommendation agent system finds that the user is not interested in the proposed target object during the interaction with the user through the recommendation assistant, and then gives up recommending the proposed target object to the user.

[0062] Optionally, the above-mentioned target object to be recommended can be pre-set into the recommendation agent system.

[0063] Optionally, the above-mentioned target object to be recommended may be target content to be recommended (such as text, pictures, audio, video, books, movies, games, online packages, etc.) or target product to be recommended (such as products in the commercial field such as goods and services). Of course, the target object to be recommended may also be other, which is not specifically limited in this application.

[0064] Optionally, the process of the memory module 53 acquiring historical human-computer interaction data and proposed target objects related to user intention may specifically include: acquiring historical human-computer interaction data from a first memory bank according to user intention; and acquiring proposed target objects from a second memory bank according to user intention.

[0065] That is, the present embodiment can pre-establish a first memory bank and a second memory bank, the first memory bank is used to store the interaction data when the user interacts with the recommendation assistant, and the second memory bank is used to store the pre-configured target objects to be recommended. Then, after determining the user's intention, the present embodiment can obtain the historical human-computer interaction data related to the user's intention from the first memory bank, and obtain the target objects to be recommended related to the user's intention from the second memory bank.

[0066] The above-mentioned planning module 52 can retrieve reference workflows that match the user's intention. Here, the reference workflow refers to the orchestration workflow used when successfully recommending objects in scenes related to the user's intention, or the manually orchestrated workflow that matches the user's intention and verifies the recommendation effect. Among them, "verifying the recommendation effect" means performing a recommendation simulation on the manually orchestrated workflow and making a successful recommendation in the recommendation simulation, and the scene related to the user's intention refers to the scene whose similarity with the user's intention is greater than the preset similarity threshold.

[0067] Optionally, the object in the above-mentioned user intention-related scene can be recommended content in the user intention-related scene (such as text, pictures, audio, video, books, movies, games, online packages, etc.), or recommended products in the user intention-related scene (such as products in the commercial field such as goods and services). Of course, the object in the user intention-related scene can also be other, which is not specifically limited in this application.

[0068] Optionally, the reference workflow includes the orchestration results of tool usage, logical judgment and model calling. Of course, the reference workflow may also include other orchestration results, which are not specifically limited in this application.

[0069] Optionally, the process of "retrieving reference workflows matching the user intention" of the planning module 52 may include: generating a query vector based on the user intention, and querying a pre-generated vector database for a reference workflow matching the user intention according to the query vector.

[0070] Specifically, the recommendation agent system will store a series of reference workflows in the vector database in advance. These workflows are encoded and processed by the vectorization model and converted into vectorized representations to support efficient retrieval and matching of the recommendation agent system. In this recommendation task, the planning module 52 analyzes the user's intention and converts it into a query vector, matches the relevant workflows in the vector database, and finds the workflow that best matches the requirements of this recommendation task from the vector database as the reference workflow that matches the retrieved user intention. This retrieval process combines semantic similarity and task requirements to ensure that the retrieved reference workflow meets the user's intention.

[0071] Furthermore, the planning module 52 can also call the recommendation model to generate an initial intelligent agent workflow based on the data obtained by the perception module, historical human-computer interaction data, the target object to be recommended and the reference workflow.

[0072] Here, the initial agent workflow is a workflow planned and arranged by the recommendation big model, using the reference workflow as a thinking chain for the recommendation big model reasoning. That is, the recommendation big model can learn the thinking mode successfully recommended by the reference workflow, and use the learned thinking mode for the current recommendation task, so that the recommendation big model can decompose the current recommendation task into multiple subtasks, arrange the multiple subtasks according to the thinking mode provided by the reference workflow, and obtain the arranged initial agent workflow.

[0073] Since the recommendation big model can draw on the thinking mode of successful recommendation of the reference workflow to generate the initial intelligent agent workflow, the accuracy of the automatic arrangement of the recommendation big model is improved, and then the accuracy of the recommendation can be improved by making recommendations based on the initial intelligent agent workflow.

[0074] In an optional embodiment, in order to further improve the accuracy of recommendation, the planning module 52 can also perform further dynamic planning on the initial agent workflow.

[0075] Specifically, in the process of calling the execution module 54 to execute the initial agent workflow, the planning module 52 can also conduct one or more rounds of dialogue with the user through the recommendation assistant. Optionally, the step-by-step dialogue guidance strategy generated by the recommendation large model can be used to conduct one or more rounds of dialogue with the user to obtain the second dialogue data. Here, the step-by-step dialogue guidance strategy can guide the user to provide more valuable information during the dialogue process, so as to more accurately understand the user's intention based on the valuable information, and then more accurately recommend relevant objects to the user.

[0076] At the same time, data may also be returned during the execution of the initial agent workflow. For example, some data may be retrieved from the recommendation database and some data may be retrieved from the Internet engine during the execution of the initial agent workflow. Then, the data retrieved from the recommendation database and the data retrieved from the Internet engine can be used as the returned data. Of course, the returned data can also be other data, which is not limited in this application.

[0077] After the planning module 52 obtains the second dialogue data and the returned data, it can determine whether to update the initial agent workflow based on the second dialogue data and / or the returned data. If so, the initial agent workflow is updated based on the first dialogue data, the second dialogue data and the reference workflow, and then the execution module 54 is called to execute the updated agent workflow from the updated position. In the process of executing the updated agent workflow, the recommendation assistant can also conduct one or more rounds of dialogue with the user again to obtain the second dialogue data. In addition, data may be returned during the execution of the initial agent workflow. In this case, it can be determined whether to update the first updated agent workflow based on the second dialogue data and the returned data. If so, the first updated agent workflow is updated for the second time, and so on, until the recommendation is completed.

[0078] In this embodiment, the above-mentioned process of updating the initial intelligent agent workflow based on the first dialogue data, the second dialogue data and the reference workflow can be completed within the recommended large model. Then, the first dialogue data and the second dialogue data can be input into the fourth prompt word prompt template of the recommended large model as dialogue contexts. At the same time, the reference workflow and the initial intelligent agent workflow can also be input into the fourth prompt word prompt template to obtain the fourth prompt instruction prompt. The fourth prompt instruction prompt is input into the recommended large model to obtain the updated intelligent agent workflow output by the recommended large model.

[0079] It should be noted that the above-mentioned "fourth prompt word prompt template" is only used to distinguish it from the following "first prompt word prompt template", "second prompt word prompt template" and "third prompt word prompt template", and is not used to limit the order of precedence; similarly, the "fourth" in the "fourth prompt instruction prompt" is only used to distinguish it from the following "first four prompt instructions prompt", "second prompt instruction prompt" and "third prompt instruction prompt", and is not used to limit the order of precedence.

[0080] In order to enable technicians in this field to better understand the above dynamic update process, an initial intelligent agent workflow that sequentially executes subtask 11, subtask 12, subtask 13, subtask 14, and subtask 15 is used as an example for explanation.

[0081] First, execute subtask 11. If after executing subtask 11, it is determined that the initial agent workflow is not to be updated, then execute subtask 12. If after executing subtask 12, it is determined that the initial agent workflow is to be updated, then update subtask 13 to subtask 23, update subtask 14 to subtask 24, and update subtask 15 to subtask 25. Then execute subtask 23. If after executing subtask 23, it is determined that an update is required, then update subtask 24 to subtask 34, and update subtask 25 to subtask 35. Then execute subtask 34 again, and determine again whether to update. If not, execute subtask 35 to obtain the recommended object and recommend it to the user.

[0082] That is, the planning module 52 can think and adjust the plan during the execution of the initial agent workflow, so that the updated agent workflow is more accurate.

[0083] In some scenarios, if it is determined not to update the initial agent workflow based on the second dialogue data and / or return data generated during the execution process, the present application can complete the recommendation based on the initial agent workflow.

[0084] Optionally, in some scenarios, as the recommendation process generates more data, user intent may change. In order to avoid recommendation errors at this time and bring a bad recommendation experience to users, this embodiment can also analyze the user intent again in combination with the second dialogue data and / or return data after the second dialogue data and / or return data is generated during the execution process, and re-query the historical human-computer interaction data, the target object to be recommended and the reference workflow based on the newly obtained user intent, regenerate and execute the initial intelligent agent workflow until the recommendation is completed.

[0085] It should be noted that the above-mentioned "completion of recommendation" can be the successful recommendation of the intended target object to the user, or it can be the abandonment of the intended target object and the recommendation of other objects of interest to the user, or the user actively ending the conversation and ending the recommendation process (the relevant object may not be successfully recommended to the user), etc.

[0086] The recommendation intelligent agent system provided by the present application comprises a perception module that obtains first dialogue data and / or multimodal environment data collected during a human-computer interaction process, determines user intent based on the obtained data, a memory module that obtains historical human-computer interaction data and a target object to be recommended related to the user intent, a planning module that retrieves a reference workflow that matches the user intent, generates an initial intelligent agent workflow based on the data obtained by the perception module, the historical human-computer interaction data, the target object to be recommended and the reference workflow, calls an execution module to execute the initial intelligent agent workflow, determines whether to update the initial intelligent agent workflow based on second dialogue data and / or return data generated during the execution process, and if so, updates the initial intelligent agent workflow based on the first dialogue data, the second dialogue data and the reference workflow, and calls the execution module to execute the updated intelligent agent workflow from the update position to complete the recommendation.

[0087] Since the first conversation data and the multimodal environment data are data generated when the user needs a recommendation, the first conversation data and the multimodal environment data can better reflect the user's current preferences, and the historical human-computer interaction data can reflect the user's past interests. Therefore, the recommendation agent system provided in the present application refers to the first conversation data, multimodal environment data and historical human-computer interaction data at the same time when making recommendations, and can more accurately determine the recommendation objects that the user prefers, thereby improving the accuracy of recommendations.

[0088] The proposed target object provides the recommendation direction for the recommendation process. The initial intelligent agent workflow is planned based on the proposed target object, and then product recommendations are made based on the initial intelligent agent workflow. This can guide users to establish interest in the proposed target object during the recommendation process, thereby increasing the probability of successful recommendation of the proposed target object. At the same time, the existence of the proposed target object makes active recommendation possible.

[0089] Furthermore, the historical human-computer interaction data in this application is the user's long-term historical memory data. Combining the historical human-computer interaction data with the target object to be recommended enables this application to plan a long-term recommendation strategy and increase the probability of successful recommendation.

[0090] Furthermore, the present application can draw on the successful recommendation experience of objects in scenes related to user intentions, and use the reference workflow used when successfully recommending the object as the thinking chain used by the recommendation large model reasoning, so that the recommendation large model can draw on the thinking mode of successful recommendation, improve the accuracy of the automatic planning of the recommendation large model, and make recommendations based on the more accurate initial intelligent agent workflow arranged by the recommendation large model, thereby improving the accuracy of the recommendation. Even if the orchestration workflow used to successfully recommend objects in scenes related to user intentions cannot be retrieved, the present application can still use the manually orchestrated workflow that matches the user intention and verifies the recommendation effect as the reference workflow, so that the initial intelligent agent workflow generated based on the reference workflow can be more effectively recommended, thereby improving the recommendation success rate.

[0091] In addition, regardless of whether the current recommendation task is complex or not, this application can retrieve the corresponding reference workflow, and generate the initial intelligent agent workflow after learning the recommendation thinking mode of the reference workflow, so that the initial intelligent agent workflow can provide a better recommendation strategy for the current recommendation task, improving the recommendation accuracy and success rate. Compared with the existing recommendation scheme that can only solve simple recommendation tasks, this application can provide an accurate initial intelligent agent workflow to solve complex recommendation tasks, and has a wider range of applications.

[0092] Furthermore, the recommendation agent system provided by the present application can adopt a step-by-step conversational guidance strategy to conduct one or more rounds of conversations with the user to obtain second conversation data during the process of making recommendations based on the automatically planned initial agent workflow. The addition of the second conversation data and the return data generated during the execution of the initial agent workflow enables the present application to better understand the user's intention, and then dynamically update the initial agent workflow based on the understanding of the user's intention, so that the updated agent workflow is more in line with the user's intention, thereby improving the accuracy of the recommendation.

[0093] The following introduces the process of generating the initial intelligent agent workflow by the planning module 52 based on the data obtained by the perception module, the historical human-computer interaction data, the target object to be recommended and the reference workflow.

[0094] In this embodiment, the recommendation model can be pre-trained, and the parameters of the recommendation model can be fine-tuned through the pre-training process, so that the recommendation model can plan and infer a more accurate initial intelligent agent workflow. Here, the training data used by the recommendation model in the training stage includes: training multimodal environment data and / or training dialogue data, training historical human-computer interaction data, training target objects to be recommended, training reference workflows, and corresponding training initial intelligent agent workflows.

[0095] Optionally, the recommendation large model pre-trained in this embodiment may be a Qwen-7B large model or a Llama-7B large model.

[0096] After training, this embodiment can fill the data acquired by the perception module, historical human-computer interaction data, the target object to be recommended and the reference workflow into the first prompt word prompt template of the recommendation model to obtain the first prompt instruction prompt, and then input the first prompt instruction prompt into the recommendation model to obtain the initial intelligent agent workflow output by the recommendation model.

[0097] Optionally, the process of "filling the data obtained by the perception module, the historical human-computer interaction data, the target object to be recommended and the reference workflow into the first prompt word prompt template of the recommendation large model" may include: obtaining the first prompt word prompt template, wherein the first prompt word prompt template includes a first basic data input slot, an interaction data input slot, an object input slot and a workflow input slot, and the first prompt word prompt template is used to instruct the recommendation large model to understand the user's intention in combination with the data in the included slots and give an initial intelligent agent workflow; filling the data obtained by the perception module into the first basic data input slot, filling the historical human-computer interaction data into the interaction data input slot, filling the target object to be recommended into the object input slot, and filling the reference workflow into the workflow input slot to obtain the first prompt instruction prompt.

[0098] This embodiment fine-tunes the recommendation big model so that the big model can improve the task planning ability of the recommendation scenario; fills the reference workflow into the preset first prompt word prompt template, and provides the background information required for the recommendation big model to infer and generate new workflows. Finally, the recommendation big model infers based on the first prompt word prompt template and the reference workflow to generate a new, customized initial intelligent agent workflow, thereby realizing intelligent planning and optimization of tasks, so that the recommendation big model can still plan a reasonable and accurate initial intelligent agent workflow when facing complex recommendation tasks, thereby improving the recommendation accuracy.

[0099] In another possible implementation, in order to avoid the recommendation big model from generating unreasonable initial intelligent agent workflows, the present application can pre-configure custom workflows and workflow generation rules, and before using the recommendation big model to generate the initial intelligent agent workflow, first train the recommendation big model based on the pre-configured custom workflows and workflow generation rules, so that the recommendation big model can follow the execution logic and execution order in the custom workflow to generate the required workflow under the guidance of the workflow generation rules.

[0100] Optionally, a custom workflow includes the following subtasks: Start: initial state; SQL: internal database search; Search: search for the internal vector library or external application programming interface (Application Programming Interface, API); Retrieve: recommend large model to perform recovery actions based on the Search results; Recall: recall operation; Rerank: reordering; Analyze: data summary analysis; Finish: task completion.

[0101] In the above custom workflow, the retrieval and search process of each subtask specifies which tools need to be used (including recommendation tools, retrieval tools, and how to read external application programming interfaces API), what logical judgments need to be made, which large models need to be called, etc. Therefore, by making recommendations according to the execution order and execution logic of these subtasks, the required data can be retrieved more quickly during the recommendation process, thereby improving the recommendation efficiency.

[0102] Optionally, specifically for this embodiment, the recommended large model "generating the required workflow following the execution logic and execution order in the custom workflow" includes generating a workflow that includes the above-mentioned initial state, task completion, and at least one subtask between the initial state and the task completion, and the workflow follows the execution logic and execution order of each subtask included.

[0103] In an embodiment of the present application, a recommendation big model is trained based on workflow generation rules and custom workflows, so that the recommendation big model can follow the execution logic and execution order in the custom workflow when planning and executing recommendation tasks, avoiding the generation of unreasonable initial intelligent agent workflows, such as generating initial intelligent agent workflows that skip intermediate retrieval and search steps, thereby improving the performance and accuracy of the recommendation big model in complex tasks.

[0104] In a possible implementation, before the memory module 53 obtains the historical human-computer interaction data related to the user intention and the target object to be recommended, the recommendation agent system provided by the present application can also determine the user intention through the perception module 51.

[0105] Specifically, the perception module 51 may also determine the user intention based on the data acquired by itself (ie, the first conversation data and / or the multimodal environment data).

[0106] Optionally, the perception module 51 can call the recommendation model to determine the user intention, that is, the perception module 51 can fill the data obtained by itself into the second prompt word prompt template of the recommendation model to obtain the second prompt instruction prompt, and input the second prompt instruction prompt into the recommendation model to obtain the user intention output by the recommendation model. Here, the training data used by the recommendation model in the training stage includes: training multimodal environment data and / or training dialogue data, and corresponding training user intentions.

[0107] Optionally, the second prompt word prompt template includes a second basic data input slot, and the second prompt word prompt template is used to instruct the recommendation model to combine the data in the second basic data input slot to infer and give the user's intention. Then, the present application can fill the data obtained by the perception module 51 into the second basic data input slot to obtain the above-mentioned second prompt instruction prompt.

[0108] In the embodiment of the present application, the powerful analytical reasoning ability of the recommendation big model is utilized to perform reasoning analysis on the user intention through the recommendation big model, thereby improving the accuracy of the user intention and further improving the recommendation accuracy.

[0109] Following the above introduction, the historical human-computer interaction data acquired by the memory module 53 can be a content summary of the historical conversation data between the user and the recommendation assistant. In the embodiment of the present application, there can be multiple methods for extracting content summaries from historical conversation data, which are provided here but not limited to the following methods.

[0110] First, keyword extraction methods, such as using keyword extraction algorithms in natural language processing technology to extract the most important words or phrases from historical conversation data. These words or phrases can represent the main content of the historical conversation data.

[0111] Second, use automatic summarization technology, such as statistical-based summarization methods, graph-based summarization methods, or deep learning-based summarization methods, to automatically generate content summaries of historical conversation data.

[0112] Third, methods using artificial intelligence or machine learning models, such as training a machine learning model or a deep learning model, output a content summary of historical conversation data based on the model.

[0113] Fourth (preferred solution), use the recommendation big model to obtain the content summary of the historical conversation data. The process may include: filling the historical conversation data into the third prompt word prompt template of the recommendation big model to obtain the third prompt instruction prompt, inputting the third prompt instruction prompt into the recommendation big model, and obtaining the content summary output by the recommendation big model; here, the training data used by the recommendation big model in the training stage includes: training conversation data and the corresponding training content summary.

[0114] Optionally, the third prompt word prompt template includes a dialogue data input slot, and the third prompt word prompt template is used to instruct the recommendation model to give a specific content summary in combination with the data in the dialogue data input slot. Then, the present application can fill the historical dialogue data into the dialogue data input slot to obtain the third prompt instruction prompt.

[0115] It should be noted that, in addition to the above method, other methods may be used to obtain a content summary of historical conversation data, which is not specifically limited in this application.

[0116] See also Figure 5 , which is an optional schematic diagram of the interaction between various modules when dynamic planning is performed based on a recommendation large model, provided in an embodiment of the present application.

[0117] like Figure 5 The perception module 51 includes a user interaction module 511 and a multimodal perception module 512. The user interaction module 511 is used to obtain the first dialogue data, and the multimodal perception module 512 (optionally, the multimodal perception module 512 can be a multimodal interaction large model) is used to obtain multimodal environment data.

[0118] The above-mentioned first dialogue data and multimodal environment data can be sent to the dynamic programming executor 521 in the planning module 52. The dynamic programming executor 521 can call the recommendation model 522 to determine the user intention based on the first dialogue data and the multimodal environment data. Then, the dynamic programming executor 521 obtains the historical human-computer interaction data related to the user intention from the first memory bank 531 in the memory module 53, obtains the target object to be recommended related to the user intention from the second memory bank 532, and obtains the reference workflow matching the user intention from the workflow library 524. After that, the dynamic programming executor 521 calls the recommendation model 522 again based on the first dialogue data, multimodal environment data, historical human-computer interaction data, target object to be recommended and reference workflow, and combines with the orchestration object 541 to generate the initial intelligent agent workflow.

[0119] After the initial intelligent agent workflow is generated, the initial intelligent agent workflow can be executed through the execution module 54. Based on the second dialogue data and / or return data generated during the execution process, it is determined whether to update the initial intelligent agent workflow. If so, the initial intelligent agent workflow is updated according to the first dialogue data, the second dialogue data and the reference workflow. If it is determined that the initial intelligent agent workflow is not to be updated, the initial intelligent agent workflow continues to be executed. Until the initial intelligent agent workflow or the updated intelligent agent workflow is executed, the recommended related objects are fed back to the user or the user ends the conversation. When the workflow evaluation module 523 determines through evaluation that the related objects are successfully recommended based on the final updated intelligent agent workflow or the initial intelligent agent workflow, the final updated intelligent agent workflow or the initial intelligent agent workflow is written into the workflow library 524.

[0120] It should be noted that there may be a situation where the initial intelligent agent workflow generated by the recommendation large model 522 is the same as the reference workflow, or there may be a situation where the final updated intelligent agent workflow is the same as the reference workflow. Since the reference workflow has been stored in the workflow library, even if the workflow evaluation module 523 determines through evaluation that the relevant objects are successfully recommended based on the final updated intelligent agent workflow or the initial intelligent agent workflow, the final updated intelligent agent workflow or the initial intelligent agent workflow will not be written to the workflow library 524.

[0121] Optionally, the above-mentioned orchestration objects include but are not limited to the following: tools used in the recommendation process, models called, and judgment logic used.

[0122] Optionally, the tools used in the recommendation process include but are not limited to: a first recall tool Recall0, a second recall tool Recalln, a reranking tool Rerank, a database retrieval tool, an API call tool, a knowledge base retrieval tool, a vector library retrieval tool, a tool for writing and executing Python scripts, a tool for filtering data using Structured Query Language (SQL), and a tool for analyzing API results.

[0123] In summary, this application flexibly arranges the orchestration objects based on the automatic planning capability of the recommendation agent system to generate an initial agent workflow, and dynamically plans and adjusts the initial agent workflow based on the continuous recommendation capability of the recommendation agent system, so that the updated agent workflow is more accurate, thereby improving the user satisfaction and recommendation success rate of conversational recommendations.

[0124] The present application also provides a recommendation method, which can be applied to the recommendation agent system described above. In order to make those skilled in the art better understand, the recommendation method of the embodiment of the present application is described in detail below in conjunction with the accompanying drawings.

[0125] Reference Figure 6 , Figure 6 A flowchart of a recommended method provided in an embodiment of the present application, the method may include: Step S601: Acquire first dialogue data and / or multimodal environment data collected during a human-computer interaction process.

[0126] The first dialogue data includes at least one round of dialogue data.

[0127] Step S602: Obtain historical human-computer interaction data related to user intention and target objects to be recommended.

[0128] Among them, user intention reflects the recommended objects that the user prefers.

[0129] Step S603: retrieve a reference workflow that matches the user's intention, and generate an initial agent workflow based on the first dialogue data and / or multimodal environment data, historical human-computer interaction data, the target object to be recommended, and the reference workflow.

[0130] Among them, the reference workflow refers to the orchestration workflow used to successfully recommend objects in scenarios related to user intentions, or the manually orchestrated workflow after matching user intentions and verifying the recommendation effect. The initial agent workflow is a workflow planned and orchestrated by the recommendation big model, using the reference workflow as a thinking chain used for reasoning of the recommendation big model.

[0131] For example, in the package recommendation scenario, the user's intention is "The current 5G package is too expensive and not cost-effective. Compared with other products on the market, I hope you can help recommend a cost-effective package." Then the generated initial intelligent agent workflow can be: query the relevant database to obtain the user's package name; call the search engine to query the price of the same product on the market; find the top n suitable packages (i.e. the top n most suitable packages) through the recommendation engine according to the user's usage behavior preferences in the user portrait; compare the cost-effectiveness of the packages of friendly companies retrieved by the search engine in turn, and select the best package; generate reply content.

[0132] Step S604, execute the initial intelligent agent workflow, and determine whether to update the initial intelligent agent workflow based on the second dialogue data and / or return data generated during the execution process. If so, update the initial intelligent agent workflow based on the first dialogue data, the second dialogue data and the reference workflow, and execute the updated intelligent agent workflow from the update position to complete the recommendation.

[0133] The second dialogue data is at least one round of dialogue data conducted with the user based on the step-by-step dialogue guidance strategy.

[0134] The recommendation method provided by the present application, since the first conversation data and the multimodal environment data are data generated when the user needs a recommendation, the first conversation data and the multimodal environment data can better reflect the user's current preferences, and the historical human-computer interaction data can reflect the user's past interests. Therefore, when making recommendations, the present application refers to the first conversation data, the multimodal environment data and the historical human-computer interaction data at the same time, which can more accurately determine the recommendation objects that the user prefers, thereby improving the recommendation accuracy.

[0135] The proposed target object provides the recommendation direction for the recommendation process. The initial intelligent agent workflow is planned based on the proposed target object, and then product recommendations are made based on the initial intelligent agent workflow. This can guide users to establish interest in the proposed target object during the recommendation process, thereby increasing the probability of successful recommendation of the proposed target object. At the same time, the existence of the proposed target object makes active recommendation possible.

[0136] Furthermore, the historical human-computer interaction data in this application is the user's long-term historical memory data. Combining the historical human-computer interaction data with the target object to be recommended enables this application to plan a long-term recommendation strategy and increase the probability of successful recommendation.

[0137] Furthermore, the present application can draw on the successful recommendation experience of objects in scenes related to user intentions, and use the reference workflow used when successfully recommending the object as the thinking chain used by the recommendation large model reasoning, so that the recommendation large model can draw on the thinking mode of successful recommendation, improve the accuracy of the automatic planning of the recommendation large model, and make recommendations based on the more accurate initial intelligent agent workflow arranged by the recommendation large model, thereby improving the accuracy of the recommendation. Even if the orchestration workflow used to successfully recommend objects in scenes related to user intentions cannot be retrieved, the present application can still use the manually orchestrated workflow that matches the user intention and verifies the recommendation effect as the reference workflow, so that the initial intelligent agent workflow generated based on the reference workflow can be more effectively recommended, thereby improving the recommendation success rate.

[0138] In addition, regardless of whether the current recommendation task is complex or not, this application can retrieve the corresponding reference workflow, and generate the initial intelligent agent workflow after learning the recommendation thinking mode of the reference workflow, so that the initial intelligent agent workflow can provide a better recommendation strategy for the current recommendation task, improving the recommendation accuracy and success rate. Compared with the existing recommendation scheme that can only solve simple recommendation tasks, this application can provide an accurate initial intelligent agent workflow to solve complex recommendation tasks, and has a wider range of applications.

[0139] Furthermore, in the process of making recommendations based on the automatically planned initial intelligent agent workflow, the present application can adopt a step-by-step conversational guidance strategy to conduct one or more rounds of conversations with the user to obtain second conversation data. The addition of the second conversation data and the return data generated during the execution of the initial intelligent agent workflow enables the present application to better understand the user's intention, and then dynamically update the initial intelligent agent workflow based on the understanding of the user's intention, so that the updated intelligent agent workflow is more in line with the user's intention, thereby improving the accuracy of the recommendation.

[0140] A recommendation method provided by an embodiment of the present application is introduced above, and a device for executing the above recommendation method will be introduced below.

[0141] See also Figure 7 , Figure 7 Schematic diagram of a recommendation device provided in an embodiment of the present application, which can be applied to the above-mentioned recommendation agent system. Figure 7 As shown, the recommended device includes: The data acquisition unit 801 is used to acquire first dialogue data and / or multimodal environment data collected during the human-computer interaction process, wherein the first dialogue data includes at least one round of dialogue data; A memory extraction unit 802 is used to obtain historical human-computer interaction data related to user intention and target objects to be recommended, wherein user intention reflects the recommended objects that the user prefers; The workflow orchestration unit 803 is used to retrieve a reference workflow matching the user's intention, and generate an initial agent workflow according to the first dialogue data and / or the multimodal environment data, the historical human-computer interaction data, the target object to be recommended, and the reference workflow, wherein the reference workflow refers to an orchestration workflow used when successfully recommending an object in a scene related to the user's intention, or a manually orchestrated workflow that matches the user's intention and verifies the recommendation effect, and the initial agent workflow is a workflow that uses the reference workflow as a thinking chain used for reasoning of the recommendation big model and is planned and orchestrated by the recommendation big model; The dynamic planning unit 804 is used to execute the initial intelligent agent workflow, and determine whether to update the initial intelligent agent workflow based on the second dialogue data and / or return data generated during the execution process. If so, the initial intelligent agent workflow is updated based on the first dialogue data, the second dialogue data and the reference workflow, and the updated intelligent agent workflow is executed from the updated position to complete the recommendation, wherein the second dialogue data is at least one round of dialogue data with the user based on the step-by-step dialogue guidance strategy.

[0142] The recommendation device provided in this application corresponds to the recommendation method and recommendation intelligent agent system provided in the previous text. For details, please refer to the previous introduction and will not be repeated here.

[0143] The present application also provides an electronic device in an embodiment. Figure 8 As shown, it shows a schematic diagram of the structure of an electronic device suitable for implementing the embodiment of the present application. The electronic device in the embodiment of the present application may include but is not limited to fixed terminals such as mobile phones, laptops, PDAs (personal digital assistants), PADs (tablet computers), desktop computers, etc. Figure 8 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0144] like Figure 8 As shown, the electronic device may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 901, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 902 or a program loaded from a storage device 908 to a random access memory (RAM) 903. When the electronic device is powered on, various programs and data required for the operation of the electronic device are also stored in the RAM 903. The processing device 901, the ROM 902, and the RAM 903 are connected to each other via a bus 904. An input / output (I / O) interface 905 is also connected to the bus 904.

[0145] Typically, the following devices may be connected to the I / O interface 905: an input device 906 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 907 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 908 including, for example, a memory card, a hard disk, etc.; and a communication device 909. The communication device 909 may allow the electronic device to communicate with other devices wirelessly or by wire to exchange data. Although Figure 8 An electronic device having various devices is shown, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0146] 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 any one of the recommendation methods provided in the embodiments of the present application.

[0147] A computer-readable storage medium is also provided in an embodiment of the present application. The storage medium carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement any recommendation method provided in the embodiment of the present application.

[0148] 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.

[0149] 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 magnetic 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.

[0150] 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.

[0151] 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. A recommendation agent system, characterized in that: include: Perception module, memory module, planning module and execution module; The perception module is used to obtain first dialogue data and / or multimodal environment data collected during the human-computer interaction process, wherein the first dialogue data includes at least one round of dialogue data; The memory module is used to obtain historical human-computer interaction data related to user intention and the target object to be recommended, wherein the user intention reflects the recommended object that the user prefers; The planning module is used to retrieve the reference workflow matching the user intention, and generate an initial agent workflow based on the data acquired by the perception module, the historical human-computer interaction data, the target object to be recommended and the reference workflow, wherein the reference workflow refers to the orchestration workflow used when the object in the scene related to the user intention is successfully recommended, or the manually orchestrated workflow after matching the user intention and verifying the recommendation effect, and the initial agent workflow is a workflow planned and orchestrated by the recommendation big model using the reference workflow as a thinking chain used for reasoning of the recommendation big model; The planning module is also used to call the execution module to execute the initial intelligent agent workflow, and determine whether to update the initial intelligent agent workflow based on the second dialogue data and / or return data generated during the execution process. If so, the initial intelligent agent workflow is updated based on the first dialogue data, the second dialogue data and the reference workflow, and the execution module is called to execute the updated intelligent agent workflow from the update position to complete the recommendation, wherein the second dialogue data is at least one round of dialogue data with the user based on a step-by-step dialogue guidance strategy.

2. The recommendation agent system according to claim 1, characterized in that: When the planning module generates the initial agent workflow according to the data acquired by the perception module, the historical human-computer interaction data, the target object to be recommended and the reference workflow, it is specifically used to: Fill the data acquired by the perception module, the historical human-computer interaction data, the proposed recommended target object and the reference workflow into the first prompt word prompt template of the recommendation large model to obtain a first prompt instruction prompt; Inputting the first prompt instruction prompt into the recommendation large model to obtain the initial agent workflow output by the recommendation large model; Among them, the training data used by the recommendation model in the training stage includes: training multimodal environment data and / or training dialogue data, training historical human-computer interaction data, training target objects to be recommended, training reference workflows, and corresponding training initial intelligent agent workflows.

3. The recommendation agent system according to claim 2, characterized in that: The training data used by the recommendation big model in the training phase also includes: preconfigured custom workflows and workflow generation rules, wherein the workflow generation rules are used to instruct the recommendation big model to follow the execution logic and execution order in the custom workflow to generate the required workflow.

4. The recommendation agent system according to claim 1, characterized in that: Before the memory module acquires the historical human-computer interaction data related to the user's intention and the target object to be recommended, the perception module is further used to: The user intention is determined based on the data acquired by the perception module.

5. The recommendation agent system according to claim 4, characterized in that: When the perception module determines the user intention according to the data acquired by the perception module, it is specifically used to: Filling the data acquired by the perception module into the second prompt word prompt template of the recommendation large model to obtain a second prompt instruction prompt; Inputting the second prompt instruction prompt into the recommendation model to obtain the user intention output by the recommendation model; The training data used by the recommendation large model in the training phase includes: training multimodal environment data and / or training conversation data, and corresponding training user intentions.

6. The recommendation agent system according to claim 1, characterized in that: When acquiring the historical human-computer interaction data and the target object to be recommended related to the user's intention, the memory module is specifically used to: According to the user intention, acquiring the historical human-computer interaction data from a first memory bank; According to the user intention, the proposed recommended target object is obtained from the second memory library.

7. The recommendation agent system according to claim 1, characterized in that: The multimodal environment data includes one or more of image data, video data, and audio data.

8. The recommendation agent system according to claim 1, characterized in that: The historical human-computer interaction data is a content summary of the historical conversation data between the user and the recommendation assistant.

9. A recommendation method, characterized in that: Applied to recommendation agent systems, including: Acquire first dialogue data and / or multimodal environment data collected during the human-computer interaction process, wherein the first dialogue data includes at least one round of dialogue data; Acquire historical human-computer interaction data related to user intention and target objects to be recommended, wherein the user intention reflects the recommended objects that the user prefers; Retrieve a reference workflow that matches the user intention, and generate an initial agent workflow based on the first dialogue data and / or the multimodal environment data, the historical human-computer interaction data, the target object to be recommended, and the reference workflow, wherein the reference workflow refers to an orchestration workflow used when successfully recommending an object in a scene related to the user intention, or a manually orchestrated workflow that matches the user intention and verifies the recommendation effect, and the initial agent workflow is a workflow that uses the reference workflow as a thinking chain used for reasoning with the recommendation big model and is planned and orchestrated by the recommendation big model; Execute the initial agent workflow, determine whether to update the initial agent workflow based on second dialogue data and / or return data generated during the execution process, and if so, update the initial agent workflow based on the first dialogue data, the second dialogue data and the reference workflow, and execute the updated agent workflow from the update position to complete the recommendation, wherein the second dialogue data is at least one round of dialogue data with the user based on a step-by-step dialogue guidance strategy.

10. A recommendation device, characterized in that: Applied to recommendation agent systems, including: A data acquisition unit, configured to acquire first dialogue data and / or multimodal environment data collected during the human-computer interaction process, wherein the first dialogue data includes at least one round of dialogue data; A memory retrieval unit, used to obtain historical human-computer interaction data related to user intention and target objects to be recommended, wherein the user intention reflects the recommended objects that the user prefers; A workflow orchestration unit, configured to retrieve a reference workflow matching the user intent, and generate an initial agent workflow according to the first dialogue data and / or the multimodal environment data, the historical human-computer interaction data, the target object to be recommended, and the reference workflow, wherein the reference workflow refers to an orchestration workflow used when successfully recommending an object in a scene related to the user intent, or a manually orchestrated workflow that matches the user intent and verifies the recommendation effect, and the initial agent workflow is a workflow planned and orchestrated by the recommendation large model using the reference workflow as a thinking chain used for reasoning of the recommendation large model; A dynamic programming unit is used to execute the initial intelligent agent workflow, determine whether to update the initial intelligent agent workflow based on the second dialogue data and / or return data generated during the execution process, and if so, update the initial intelligent agent workflow based on the first dialogue data, the second dialogue data and the reference workflow, and execute the updated intelligent agent workflow from the updated position to complete the recommendation, wherein the second dialogue data is at least one round of dialogue data with the user based on a step-by-step dialogue guidance strategy.

Images