Event dialogue collecting and processing method and system

Abstract

The invention relates to the technical field of intelligent event processing, and provides an event dialogue collecting and processing method and system.The method comprises the steps that a corresponding cache instance is created according to a received dialogue request, and the cache instance is initialized; modifying a preset information extraction cue word template according to the cache instance, obtaining an information extraction cue word, and updating slot position information of the cache instance according to the information extraction cue word; when a picture is received in the dialogue, the cache instance is updated by analyzing and adding picture information, and when an instruction of deleting the picture is received in the dialogue, the picture is deleted by modifying the picture information in the cache instance; and modifying a preset content reply prompt word template according to the cache instance, obtaining a content reply prompt word, and generating dialogue reply content according to the content reply prompt word. Through accurate cache management, slot state control and picture content maintenance, the information acquisition and processing flow is optimized, the response speed and the data processing precision of the system are improved, and the intelligent level of user interaction is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of event intelligent processing, and in particular to a method and system for event dialogue collection and processing. Background Art

[0002] With the continuous development of artificial intelligence and natural language processing technology, natural language-based dialogue systems have been widely used in various fields. These dialogue systems usually understand user needs by extracting natural language information input by users and respond accordingly. However, existing dialogue systems still face some technical challenges when processing events, especially when dealing with complex dialogue scenarios.

[0003] Most current technologies rely on natural language processing (NLP) models for information extraction, that is, semantic analysis of natural language input by users to identify key entities, actions, and events. However, this natural language-based dialogue processing method has the following major problems: 1. Lack of contextual relevance: When processing a single dialogue round, existing systems often cannot effectively use previous text or context information for reasonable reasoning, resulting in poor accuracy in information extraction and poor consistency in event processing. Especially in the case of long-term interactions or multiple rounds of dialogue, the system can easily lose the continuity of context, resulting in deviations in understanding.

[0004] 2. Ambiguity: Natural language is inherently ambiguous, especially when there is not enough contextual information. Some expressions may be misunderstood or interpreted as having multiple different meanings. For example, some vague or ambiguous words or sentences may have different interpretations depending on the context. Existing technologies have not been able to effectively solve this problem, resulting in ambiguity in the system's understanding of user intent.

[0005] 3. Misunderstanding of expressions: Since the dialogue system lacks the ability to adapt to a variety of expressions when contextual information is insufficient, the system often cannot correctly interpret the user's intentions. In particular, misunderstandings may occur when the user's expressions are irregular, brief, or ambiguous, affecting the accuracy of event processing.

[0006] Therefore, how to improve the accuracy of event processing and semantic understanding when dealing with complex conversations and multi-round interactions, and enhance the overall performance and user experience of the conversation system, has become a technical problem that needs to be solved urgently. Summary of the invention

[0007] In view of this, in order to overcome the deficiencies of the prior art, the present invention aims to provide a method and system for event conversation collection and processing.

[0008] According to a first aspect of the present invention, a method for collecting and processing event conversations is provided, the method comprising: Create a corresponding cache instance according to the received conversation request and initialize the cache instance; Modify the preset information extraction prompt word template according to the cache instance, obtain the information extraction prompt word, and update the slot information of the cache instance according to the information extraction prompt word; When a picture is received in a conversation, the cache instance is updated by parsing and adding picture information. When a command to delete a picture is received in a conversation, the picture is deleted by modifying the picture information in the cache instance. Modify the preset content reply prompt word template according to the cache instance, obtain the content reply prompt word, and generate the dialogue reply content according to the content reply prompt word.

[0009] Preferably, in the event conversation collection and processing method of the present invention, a corresponding cache instance is created according to a received conversation request, including: receiving a conversation request from a client, retrieving a corresponding cache instance from a cache system according to a session ID in the conversation request, and when a corresponding cache instance does not exist in the cache system, creating a corresponding cache instance for the conversation request, wherein the cache instance includes slot information, conversation history information and picture set information.

[0010] Preferably, in the event conversation collection and processing method of the present invention, creating a corresponding cache instance for a conversation request includes: Create slot information for a cache instance by configuring a key-value pair, where the key of the slot information is a field name, the value of the slot information is a slot structure, and the slot structure includes field description information and field current value information; Create conversation history information for the cache instance by configuring a history information list, wherein the history information list includes roles and history content, wherein the role value is user and system, and the history content value is user request and system response; The picture set information is created for the cache instance by configuring the picture set information list, wherein the picture set information list includes a picture file information object, and the picture file information object includes a picture storage address, a file sequence number, an event type after picture parsing, an event description, and a deletion mark.

[0011] Preferably, in the event conversation collection and processing method of the present invention, initializing the cache instance includes: Query the preset configuration information according to the type of the dialog request, obtain the corresponding field information from the preset configuration information, and initialize the slot information of the cache instance to empty according to the obtained field information; Initialize the historical information of the cache instance by setting the historical information list to empty; Initialize the picture set information of the cache instance by setting the picture set information list to empty.

[0012] Preferably, in the event dialogue collection and processing method of the present invention, modifying the preset information extraction prompt word template according to the cache instance to obtain the information extraction prompt word includes: Configuring an information extraction prompt word template, wherein the template parameters of the information extraction prompt word template include unknown information, unknown optional information, known information, updateable known information, current event, and known picture information; Traverse the field items of the slot information in the cache instance, and splice the field items into the template parameters corresponding to the information extraction prompt word template according to the field description information and the field current value information of the slot information; Traverse the picture set information in the cache instance, and add the valid picture information to the template parameter corresponding to the information extraction prompt word template according to the picture file information object; The information extraction prompt word template is used to generate information extraction prompt words according to the modified template parameters.

[0013] Preferably, in the event conversation collection and processing method of the present invention, the slot information of the cache instance is updated according to the information extraction prompt word, including: using the information extraction prompt word as the prompt word of the information extraction large model, using the historical conversation information in the cache instance as the input of the information extraction large model, extracting the current unknown information and the updateable known information, traversing the extracted values, and updating the slot information of the cache instance by replacing the current value of the corresponding field in the slot information with the latest extracted value.

[0014] Preferably, in the event conversation collection and processing method of the present invention, when a picture is received in the conversation, the cache instance is updated by parsing and adding picture information, and when an instruction to delete the picture is received in the conversation, the picture is deleted by modifying the picture information in the cache instance, including: When a picture is received in a conversation, the received picture is parsed to obtain picture information, wherein the picture information includes the picture storage address, file sequence number, event type after picture parsing, and event description; Add the parsed image information and the deletion mark to the image set information list of the cache instance; When a command to delete a picture is received in the conversation, the deletion mark in the picture information of the corresponding file number is set to deleted.

[0015] Preferably, in the event dialogue collection and processing method of the present invention, modifying the preset content reply prompt word template according to the cache instance to obtain the content reply prompt word includes: Configuring a content reply prompt word template, wherein the template parameters of the content reply prompt word template include unknown information, unknown optional information, known information, updateable known information, current event, and known picture information; Traverse the field items of the slot information in the cache instance, and splice the field items into the template parameters corresponding to the content reply prompt word template according to the field description information and the field current value information of the slot information; Traverse the picture set information in the cache instance, and add the valid picture information to the template parameter corresponding to the content reply prompt word template according to the picture file information object; The content reply prompt word template is used to generate content reply prompt words according to the modified template parameters.

[0016] Preferably, in the event conversation collection and processing method of the present invention, conversation reply content is generated according to content reply prompt words, including: using content reply prompt words as prompt words for generating a large model in response, using historical conversation information in the cache instance as input for generating the large model in response, and generating a request for further obtaining unknown information or a reply summarizing currently known information.

[0017] According to a second aspect of the present invention, an event conversation collection and processing system is provided, the system includes a conversation collection and processing server, the conversation collection and processing server is used to: create a corresponding cache instance according to a received conversation request, and initialize the cache instance; modify a preset information extraction prompt word template according to the cache instance, obtain the information extraction prompt word, and update the slot information of the cache instance according to the information extraction prompt word; when a picture is received in the conversation, the cache instance is updated by parsing and adding the picture information, and when an instruction to delete the picture is received in the conversation, the picture is deleted by modifying the picture information in the cache instance; modify a preset content reply prompt word template according to the cache instance, obtain the content reply prompt word, and generate conversation reply content according to the content reply prompt word.

[0018] According to a third aspect of the present invention, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in the first aspect of the present invention when executing the program.

[0019] The event dialogue collection and processing method and system of the present invention optimizes the information collection and processing process through precise cache management, slot status control and image content maintenance, improves the system response speed and data processing accuracy, and enhances the intelligent level of interaction with users, and has the following beneficial technical effects: 1. Optimize intermediate data cache: significantly improve the data storage and processing efficiency during the information collection process. By designing an optimized data cache mechanism, the system's dependence on information flow is effectively reduced, making information collection faster and more accurate, thereby improving the overall system performance.

[0020] 2. Accurately control slot storage and update status: It can accurately track and maintain the status information of each slot during the information extraction and response process, so that when the system obtains user intentions and extracts valid information, it can perform more accurate analysis and reasoning based on the stored slot metadata, thereby improving the accuracy of information extraction and reducing the risk of misidentification, helping to ask users questions and obtain more accurate information, and improving the user interaction experience.

[0021] 3. Enhance image validity and content maintenance: It can effectively verify the accuracy of image content and optimize the application of images in the information collection process. Through accurate image validity judgment and content update, it can ensure the real-time and relevance of image data, thereby improving the quality of information processing and the reliability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0023] Figure 1 This is a diagram showing an example of the architecture of an event conversation collection and processing system according to an embodiment of the present invention; Figure 2 A flowchart of a method for collecting and processing event conversations according to an embodiment of the present invention; Figure 3 A schematic diagram of the structure of the device provided by the present invention. DETAILED DESCRIPTION

[0024] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0025] It should be noted that the following embodiments and features in the embodiments may be combined with each other in the absence of conflict; and, based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in the field without making any creative work are within the scope of protection of the present disclosure.

[0026] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein may be embodied in a wide variety of forms, and any specific structure and / or function described herein is merely illustrative. Based on the present disclosure, it should be understood by those skilled in the art that an aspect described herein may be implemented independently of any other aspect, and two or more of these aspects may be combined in various ways. For example, any number of aspects described herein may be used to implement the device and / or practice the method. In addition, other structures and / or functionalities other than one or more of the aspects described herein may be used to implement this device and / or practice this method.

[0027] Figure 1 FIG. 1 is a diagram showing an example of the architecture of an event conversation collection and processing system according to an embodiment of the present invention. Figure 1 As shown, the system may include a conversation collection and processing server 101, a communication network 102 and / or one or more conversation collection and processing clients 103. Figure 1 The example shown in FIG. 1 is a plurality of conversation collection and processing clients 103 .

[0028] The conversation collection processing server 101 can be any appropriate server for storing information, data, programs and / or any other suitable type of content. In some embodiments, the conversation collection processing server 101 can perform appropriate functions. For example, in some embodiments, the conversation collection processing server 101 can be used to: create a corresponding cache instance according to the received conversation request and initialize the cache instance; modify the preset information extraction prompt word template according to the cache instance, obtain the information extraction prompt word, and update the slot information of the cache instance according to the information extraction prompt word; when a picture is received in the conversation, update the cache instance by parsing and adding the picture information, and when an instruction to delete the picture is received in the conversation, delete the picture by modifying the picture information in the cache instance; modify the preset content reply prompt word template according to the cache instance, obtain the content reply prompt word, and generate the conversation reply content according to the content reply prompt word.

[0029] As another example, in some embodiments, the conversation collection and processing server 101 may send the event conversation collection and processing method to the conversation collection and processing client 103 for user use based on the request of the conversation collection and processing client 103 .

[0030] As an optional example, in some embodiments, the conversation collection and processing client 103 is used to provide a visual conversation collection and processing interface, which is used to receive a selection input operation of a user event conversation collection and processing, and, in response to the selection input operation, obtain the conversation collection and processing interface corresponding to the option selected by the selection input operation from the conversation collection and processing server 101 and display the conversation collection and processing interface, wherein the conversation collection and processing interface at least displays information on event conversation collection and processing and operation options for the information on event conversation collection and processing.

[0031] In some embodiments, the communication network 102 can be any suitable combination of one or more wired and / or wireless networks. For example, the communication network 102 can include any one or more of the following: the Internet, an intranet, a wide area network (WAN), a local area network (LAN), a wireless network, a digital subscriber line (DSL) network, a frame relay network, an asynchronous transfer mode (ATM) network, a virtual private network (VPN), and / or any other suitable communication network. The conversation acquisition processing client 103 can be connected to the communication network 102 through one or more communication links (e.g., communication link 104), and the communication network 102 can be linked to the conversation acquisition processing server 101 via one or more communication links (e.g., communication link 105). The communication link can be any communication link suitable for transmitting data between the conversation acquisition processing client 103 and the conversation acquisition processing server 101, such as a network link, a dial-up link, a wireless link, a hard-wired link, any other suitable communication link, or any suitable combination of such links.

[0032] The conversation collection processing client 103 may include any one or more clients that present an interface related to the event conversation collection processing in an appropriate form for use and operation by a user. In some embodiments, the conversation collection processing client 103 may include any suitable type of device. For example, in some embodiments, the conversation collection processing client 103 may include a mobile device, a tablet computer, a laptop computer, a desktop computer, and / or any other suitable type of client device.

[0033] Although the conversation collection processing server 101 is illustrated as one device, in some embodiments, any appropriate number of devices may be used to perform the functions performed by the conversation collection processing server 101. For example, in some embodiments, multiple devices may be used to implement the functions performed by the conversation collection processing server 101. Alternatively, a cloud service may be used to implement the functions of the conversation collection processing server 101.

[0034] Based on the above system, an embodiment of the present invention provides an event dialogue collection and processing method, which is described below through the following embodiments.

[0035] Figure 2 FIG. 1 is a flowchart of a method for collecting and processing event conversations according to an embodiment of the present invention. The method for collecting and processing event conversations according to this embodiment can be executed at a conversation collection and processing server end, such as Figure 2 As shown, the event dialogue collection and processing method includes the following steps: Step S201: Create a corresponding cache instance according to the received dialog request and initialize the cache instance.

[0036] This embodiment receives a conversation request from a client, retrieves a corresponding cache instance from a cache system according to a session ID in the conversation request, and creates a corresponding cache instance for the conversation request when the corresponding cache instance does not exist in the cache system. The cache instance includes slot information, conversation history information, and picture collection information. The client in this embodiment includes a browser or a mobile terminal, and can also be a client that initiates a conversation with another user, which is not limited in this embodiment.

[0037] As an optional example, this embodiment creates slot information for a cache instance by configuring a key-value pair, where the key of the slot information is a field name, and the value of the slot information is a slot structure, which includes field description information and field current value information; creates conversation history information for a cache instance by configuring a history information list, where the history information list includes roles and history content, where the values ​​of the roles are users and systems, and the values ​​of the history content are user requests and system responses; creates picture set information for a cache instance by configuring a picture set information list, where the picture set information list includes a picture file information object, and the picture file information object includes a picture storage address, a file serial number, an event type after picture parsing, an event description, and a deletion mark.

[0038] For example, in this embodiment, according to the session id carried in the request, it is retrieved whether the cache exists. When the corresponding cache instance exists in the cache system, the query configuration is used to obtain the field information required for this session. The cache system is implemented in memory as a storage system with the session id as the key and the specific cache instance as the value.

[0039] When the cache instance corresponding to the current session id does not exist, create a cache instance for this session. The structure of the cache instance includes: Slot information: the field information to be obtained and its current value, implemented as a map key-value pair structure, each key-value pair is an item of information to be obtained. Conversation history information: used to store the user's request content and the current system response content during the interaction process. Image set information: used to cache the transmitted images and file information in the session context.

[0040] Specifically, the definition of the slot information Slot structure is as follows: the slot information is a dictionary object, in which the key is the field name and the value is a structure, including: the description information itemschema of the field and the currently obtained value information itemvalue.

[0041] The definition of the conversation history information history structure is as follows: It is used to store the user request content and the current system response content during the user interaction process. It is defined as a list of {role, content}, where the role values ​​are user and assistant, representing the user request and system response respectively. It is implemented as: [{role:user,content:user request},{role:assistant,content:system response content},...]. In this embodiment, the system-level prompt words are not stored. The system-level prompt words are customized by the reply response link each time they are used.

[0042] The structure of the picture set information imgset is defined as follows: The picture set information imgset is a list, which contains a series of uploaded and parsed picture file information objects, including the picture storage address imgurl, the file sequence number imgSeq, the event type eventType and event description desc after the picture is parsed, and the deletion mark isDelete. The deletion mark is used as the basis for recording file deletion and content analysis.

[0043] After the cache instance is created, the present embodiment initializes the created cache instance. As an optional example, the present embodiment queries the preset configuration information according to the type of the dialog request, obtains the corresponding field information from the preset configuration information, and initializes the slot information of the cache instance to empty according to the obtained field information; initializes the historical information of the cache instance by setting the historical information list to empty; and initializes the picture set information of the cache instance by setting the picture set information list to empty.

[0044] For example, this embodiment queries the configuration to obtain the field information required for this session, initializes the schema of the cache slot according to the required field information, and sets the value to empty. The schema includes whether it is required, whether it can be updated, and the default value. In the slot information Slot, the field information required for this session is not known when the cache instance is initialized. This step queries the preset configuration information according to the current session type to obtain the field information to be obtained. For example, event time, reporter, contact number, etc. For each item to be obtained, the configuration configures its field name name, English name nameEn, default value defaultVal, field description itemdesc, whether it is required isRequired, and field restriction remark. This embodiment initializes the slot information of the cache instance in sequence based on this configuration information. Specifically, according to the obtained configuration information, an object with the key as the field name and the value as a structure is inserted into the slot information Slot in sequence, where the structure includes: the description information itemschema of the field and the currently obtained value information itemvalue. The description information itemschema of this field describes the specific definition of this field and the information to be collected, including the configured field name name, English name nameEn, default value defaultVal, field description itemdesc, whether it is required isRequired, and field restriction remark. The current value itemvalue of this field defaults to empty, and the default value set in itemschema is finally used to assign a value to itemvalue.

[0045] After the slot information is initialized, this embodiment initializes the conversation history and picture information to be empty, initializes the list of picture set information imgset to be empty, returns the cache instance object created this time, and all subsequent multi-step operations of this session are mediated by this cache instance object.

[0046] Step S202: modifying a preset information extraction prompt word template according to the cache instance, obtaining information extraction prompt words, and updating the slot information of the cache instance according to the information extraction prompt words.

[0047] As an optional example, this embodiment configures an information extraction prompt word template, and the template parameters of the information extraction prompt word template include unknown information, unknown optional information, known information, updateable known information, current events, and known image information; traverses the field items of the slot information in the cache instance, and splices the field items into the template parameters corresponding to the information extraction prompt word template according to the field description information of the slot information and the field current value information; traverses the picture set information in the cache instance, and adds the valid picture information to the template parameters corresponding to the information extraction prompt word template according to the picture file information object; uses the information extraction prompt word template to generate information extraction prompt words according to the modified template parameters.

[0048] After generating the information extraction prompt word, this embodiment uses the information extraction prompt word as the prompt word of the information extraction large model, uses the historical conversation information in the cache instance as the input of the information extraction large model, extracts the current unknown information and updateable known information, traverses the extracted values, and updates the slot information of the cache instance by replacing the current value of the corresponding field in the slot information with the latest extracted value.

[0049] For example, this embodiment presets an information extraction prompt word template, and the template contains the following template parameters: unknown information, unknown optional information, known information, updateable known information, current time, known image information, etc., which need to be obtained according to the context in the implementation. Traverse each field item in the slot information Slot of the cache instance, splice the current value that is empty and the required field in the schema to [Unknown Information], splice the current value that is empty and the non-required field in the schema to [Unknown Optional Information], splice the field whose current value is not empty to [Known Information], and splice the current value that is not empty and updateable in the schema to [Updateable Known Information]. If any of the above items is empty, set the corresponding information to "None". Traverse the image set imgset, append the event type and image description that have not been deleted, the url is not empty, and the type is valid to [Known Image Information], and replace the template parameters of the information extraction prompt word template with the above values ​​to obtain the information extraction prompt word. The information extraction prompt words obtained above are used as the system-level prompt words of the large model, combined with the user's historical conversation information as the large model history information for calling, and the currently unknown and updateable known information are extracted to obtain part or all of the extraction results. The above-obtained extraction results are traversed, and the current value itemvalue of the corresponding field in the slot information Slot is replaced by the latest extraction value and set to the extracted result, and the slot information of the cache instance is updated.

[0050] Step S203: When a picture is received in the conversation, the cache instance is updated by parsing and adding picture information. When an instruction to delete the picture is received in the conversation, the picture is deleted by modifying the picture information in the cache instance.

[0051] In this embodiment, when a picture is received in a conversation, the received picture is parsed to obtain picture information, which includes the picture storage address, file number, event type after picture parsing, and event description; the picture information obtained by parsing and a deletion mark are added to the picture set information list of the cache instance; when an instruction to delete the picture is received in the conversation, the deletion mark in the picture information corresponding to the file number is set to deleted.

[0052] For example, in the conversation of this embodiment, the received picture can be analyzed by the multimodal large model to obtain the event type and picture description information in the picture, and the picture is stored in the file system, where the received picture information includes event type, picture URL, picture description and other information. The picture information is sequentially appended to imgset, and each record includes: sequence number, whether to delete, event type, picture URL and picture description. Here, a deletion mark is added, and the default is not deleted.

[0053] When receiving an instruction to delete a picture with a specified serial number, the picture record is retained, and the record with the corresponding serial number is simply set as deleted. When the picture information is used subsequently, only the valid picture information that has not been deleted is taken.

[0054] Step S204: modifying the preset content reply prompt word template according to the cache instance, obtaining the content reply prompt word, and generating the dialogue reply content according to the content reply prompt word.

[0055] As an optional example, this embodiment configures a content reply prompt word template, and the template parameters of the content reply prompt word template include unknown information, unknown optional information, known information, updateable known information, current events, and known image information; traverses the field items of the slot information in the cache instance, and splices the field items into the template parameters corresponding to the content reply prompt word template according to the field description information of the slot information and the field current value information; traverses the picture set information in the cache instance, and adds valid picture information to the template parameters corresponding to the content reply prompt word template according to the picture file information object; uses the content reply prompt word template to generate content reply prompt words according to the modified template parameters.

[0056] After generating the content reply prompt words, this embodiment uses the content reply prompt words as prompt words for generating a large model in response, uses the historical conversation information in the cache instance as input for generating the large model in response, and generates a request for further obtaining unknown information or a reply summarizing the currently known information.

[0057] For example, this embodiment presets a content reply prompt word template, and the template includes the following template parameters: unknown information, known information, updateable known information, current time, known image information, etc., which need to be obtained in the context and cache information. The current value of the current slot information Slot of the cache instance is traversed and processed, and the field name whose current value is empty and required in the schema is spliced ​​to [Unknown Information], the field name whose current value is not empty is spliced ​​to [Known Information], and the field name whose current value is not empty and updateable in the schema is spliced ​​to [Updateable Known Information]. The image set imgset is traversed and processed, and the event type and image description that are not deleted, the url is not empty, and the type is valid are appended to the known image information. The above-obtained values ​​are replaced with the template parameters of the content reply prompt word template to obtain the content reply prompt word, and the items with the above-mentioned information as empty strings are set to "None". The content reply prompt word is used as a large model system-level prompt word, combined with the user's historical dialogue information as the large model historical information for calling, and the request reply content based on the current information and unknown information is obtained. This reply is used to further obtain unknown information requests or summarize all currently known information.

[0058] The event dialogue collection and processing method and system of the embodiment of the present invention optimizes the information collection and processing process through accurate cache management, slot status control and image content maintenance, improves the system response speed and data processing accuracy, and enhances the intelligent level of interaction with users, and has the following beneficial technical effects: 1. Optimize intermediate data cache: significantly improve the data storage and processing efficiency during the information collection process. By designing an optimized data cache mechanism, the system's dependence on information flow is effectively reduced, making information collection faster and more accurate, thereby improving the overall system performance.

[0059] 2. Accurately control slot storage and update status: It can accurately track and maintain the status information of each slot during the information extraction and response process, so that when the system obtains user intentions and extracts valid information, it can perform more accurate analysis and reasoning based on the stored slot metadata, thereby improving the accuracy of information extraction and reducing the risk of misidentification, helping to ask users questions and obtain more accurate information, and improving the user interaction experience.

[0060] 3. Enhance image validity and content maintenance: It can effectively verify the accuracy of image content and optimize the application of images in the information collection process. Through accurate image validity judgment and content update, it can ensure the real-time and relevance of image data, thereby improving the quality of information processing and the reliability of the system.

[0061] like Figure 3As shown, the present invention also provides a device, including a processor 310, a communication interface 320, a memory 330 for storing a computer program executable by the processor, and a communication bus 340. The processor 310, the communication interface 320, and the memory 330 communicate with each other through the communication bus 340. The processor 310 implements the above-mentioned event dialogue collection and processing method by running the executable computer program.

[0062] Among them, the computer program in the memory 330 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0063] The system embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, i.e., they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected based on actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.

[0064] Through the description of the above implementation modes, those skilled in the art can clearly understand that each implementation mode can be implemented by means of software plus a necessary general hardware platform, or of course by hardware. Based on such an understanding, the above technical solution can essentially or in other words be embodied in the form of a software product that contributes to the prior art. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of each embodiment or some parts of the embodiment.

[0065] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A method for collecting and processing event dialogues, characterized in that: The method comprises: Create a corresponding cache instance according to the received conversation request and initialize the cache instance; Modify the preset information extraction prompt word template according to the cache instance, obtain the information extraction prompt word, and update the slot information of the cache instance according to the information extraction prompt word; When a picture is received in a conversation, the cache instance is updated by parsing and adding picture information. When a command to delete a picture is received in a conversation, the picture is deleted by modifying the picture information in the cache instance. Modify the preset content reply prompt word template according to the cache instance, obtain the content reply prompt word, and generate the dialogue reply content according to the content reply prompt word.

2. The event dialogue collection and processing method according to claim 1, characterized in that: Creating a corresponding cache instance according to a received conversation request includes: receiving a conversation request from a client, retrieving a corresponding cache instance from a cache system according to a session ID in the conversation request, and when the corresponding cache instance does not exist in the cache system, creating a corresponding cache instance for the conversation request, wherein the cache instance includes slot information, conversation history information, and picture set information.

3. The event dialogue collection and processing method according to claim 2, characterized in that: Create a corresponding cache instance for the conversation request, including: Create slot information for a cache instance by configuring a key-value pair, where the key of the slot information is a field name, the value of the slot information is a slot structure, and the slot structure includes field description information and field current value information; Create conversation history information for the cache instance by configuring a history information list, wherein the history information list includes roles and history content, wherein the role value is user and system, and the history content value is user request and system response; The picture set information is created for the cache instance by configuring the picture set information list, wherein the picture set information list includes a picture file information object, and the picture file information object includes a picture storage address, a file sequence number, an event type after picture parsing, an event description, and a deletion mark.

4. The event dialogue collection and processing method according to claim 1, characterized in that: Initialize the cache instance, including: Query the preset configuration information according to the type of the dialog request, obtain the corresponding field information from the preset configuration information, and initialize the slot information of the cache instance to empty according to the obtained field information; Initialize the historical information of the cache instance by setting the historical information list to empty; Initialize the picture set information of the cache instance by setting the picture set information list to empty.

5. The event dialogue collection and processing method according to claim 1, characterized in that: Modify the preset information extraction prompt word template according to the cache instance to obtain the information extraction prompt word, including: Configuring an information extraction prompt word template, wherein the template parameters of the information extraction prompt word template include unknown information, unknown optional information, known information, updateable known information, current event, and known picture information; Traverse the field items of the slot information in the cache instance, and splice the field items into the template parameters corresponding to the information extraction prompt word template according to the field description information and the field current value information of the slot information; Traverse the picture set information in the cache instance, and add the valid picture information to the template parameter corresponding to the information extraction prompt word template according to the picture file information object; The information extraction prompt word template is used to generate information extraction prompt words according to the modified template parameters.

6. The event conversation collection and processing method according to claim 1, characterized in that: The slot information of the cache instance is updated according to the information extraction prompt word, including: using the information extraction prompt word as the prompt word of the information extraction large model, using the historical conversation information in the cache instance as the input of the information extraction large model, extracting the current unknown information and the updateable known information, traversing the extracted values, and updating the slot information of the cache instance by replacing the current value of the corresponding field in the slot information with the latest extracted value.

7. The event dialogue collection and processing method according to claim 1, characterized in that: When a picture is received in a conversation, the cache instance is updated by parsing and adding picture information. When a command to delete a picture is received in a conversation, the picture is deleted by modifying the picture information in the cache instance, including: When a picture is received in a conversation, the received picture is parsed to obtain picture information, wherein the picture information includes the picture storage address, file sequence number, event type after picture parsing, and event description; Add the parsed image information and the deletion mark to the image set information list of the cache instance; When a command to delete a picture is received in the conversation, the deletion mark in the picture information of the corresponding file number is set to deleted.

8. The event dialogue collection and processing method according to claim 1, characterized in that: Modify the preset content reply prompt word template according to the cache instance and obtain the content reply prompt word, including: Configuring a content reply prompt word template, wherein the template parameters of the content reply prompt word template include unknown information, unknown optional information, known information, updateable known information, current event, and known picture information; Traverse the field items of the slot information in the cache instance, and splice the field items into the template parameters corresponding to the content reply prompt word template according to the field description information and the field current value information of the slot information; Traverse the picture set information in the cache instance, and add the valid picture information to the template parameter corresponding to the content reply prompt word template according to the picture file information object; The content reply prompt word template is used to generate content reply prompt words according to the modified template parameters.

9. The event conversation collection and processing method according to claim 1, characterized in that: Generate conversation reply content according to content reply prompt words, including: using content reply prompt words as prompt words for generating a large model for response, using historical conversation information in a cache instance as input for generating the large model for response, and generating a request for further obtaining unknown information or a reply summarizing currently known information.

10. An event dialogue collection and processing system, characterized in that: The system includes a conversation collection and processing server, which is used to: create a corresponding cache instance according to a received conversation request and initialize the cache instance; modify a preset information extraction prompt word template according to the cache instance, obtain the information extraction prompt word, and update the slot information of the cache instance according to the information extraction prompt word; When a picture is received in the conversation, the cache instance is updated by parsing and adding the picture information. When an instruction to delete the picture is received in the conversation, the picture is deleted by modifying the picture information in the cache instance. The preset content reply prompt word template is modified according to the cache instance, the content reply prompt words are obtained, and the conversation reply content is generated according to the content reply prompt words.

Images