Conference task processing method and device

Abstract

The invention discloses a conference task processing method and device, and belongs to the technical field of electronic equipment. The method comprises the steps of determining at least one piece of first conference task information of a target conference according to a first conference record of the target conference; generating a first task atlas of the target conference according to the at least one piece of first conference task information; wherein the first task atlas is used for performing graph structured representation on at least one piece of first conference task information in the form of task nodes and edges, and one piece of first conference task information at least comprises at least one of the following items: task content, task processor information and task deadline information.

Description

Technical Field

[0001] The present application belongs to the technical field of electronic equipment, and specifically relates to a conference task processing method and device thereof. Background Art

[0002] As modern businesses and organizations grow, meetings play a vital role in team collaboration and decision-making.

[0003] In the related art, voice recognition is usually performed on the audio files of the meeting to obtain the meeting minutes. However, since online meetings are usually long and the meeting content is complicated and rich, the meeting minutes generated by voice recognition are long and complicated. When users view the meeting minutes later, they need to spend a certain amount of time to find the content they are interested in in the meeting content, which results in low efficiency in finding the meeting content. Summary of the invention

[0004] The purpose of the embodiments of the present application is to provide a conference task processing method and device thereof, which can simplify the subsequent process of viewing conference content, enable users to quickly locate the conference tasks they are concerned about, and improve the efficiency of searching for conference content.

[0005] In the first aspect, an embodiment of the present application provides a conference task processing method, the method comprising: determining at least one first conference task information of the target meeting based on the first meeting record of the target meeting; generating a first task graph of the target meeting based on the at least one first conference task information; wherein the first task graph is used to represent at least one first conference task information in a graph structured manner in the form of task nodes and edges, and one first conference task information includes at least one of the following: task content, task handler information, and task deadline information.

[0006] In the second aspect, an embodiment of the present application provides a conference task processing device, which includes: a processing module, the processing module is used to: determine at least one first conference task information of the target meeting based on the first meeting record of the target meeting; generate a first task graph of the target meeting based on the at least one first meeting task information; wherein the first task graph is used to represent at least one first meeting task information in a graph structured manner in the form of task nodes and edges, and a first meeting task information includes at least one of the following: task content, task handler information, and task deadline information.

[0007] In a third aspect, an embodiment of the present application provides an electronic device, which includes a processor and a memory, wherein the memory stores programs or instructions that can be run on the processor, and when the program or instructions are executed by the processor, the steps of the method described in the first aspect are implemented.

[0008] In a fourth aspect, an embodiment of the present application provides a readable storage medium, on which a program or instruction is stored, and when the program or instruction is executed by a processor, the steps of the method described in the first aspect are implemented.

[0009] In a fifth aspect, an embodiment of the present application provides a chip, comprising a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run a program or instruction to implement the method described in the first aspect.

[0010] In a sixth aspect, an embodiment of the present application provides a computer program product, which is stored in a storage medium and is executed by at least one processor to implement the method described in the first aspect.

[0011] In an embodiment of the present application, the electronic device determines at least one first meeting task information of the target meeting based on the first meeting record of the target meeting, and generates a task map of the target meeting based on the at least one first meeting task information, wherein the above-mentioned task map is used to represent the at least one first meeting task information in a graph structured manner in the form of task nodes and edges, and a first meeting task information includes at least one of the following: task content, task handler information, and task deadline information. Through this method, the electronic device can extract specific meeting task information based on the meeting record, and generate a task map in the form of task nodes and edges, so that the user can intuitively view the various tasks in the meeting through the task map, as well as the relationship between the task content, task handler, and task deadline of each task. In this way, the subsequent process of viewing the content of the meeting is simplified, allowing users to quickly locate the meeting tasks they are concerned about, thereby improving the efficiency of searching the content of the meeting. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 A flowchart of a conference task processing method provided in an embodiment of the present application;

[0013] Figure 2 One of the schematic diagrams of the task map provided in the embodiment of the present application;

[0014] Figure 3 The second schematic diagram of the task map provided in the embodiment of the present application;

[0015] Figure 4 A schematic diagram of the structure of a conference task processing device provided in an embodiment of the present application;

[0016] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application;

[0017] Figure 6 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present application to clearly describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments in the present application belong to the scope of protection of this application.

[0019] The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the terms used in this way are interchangeable where appropriate, so that the embodiments of the present application can be implemented in an order other than those illustrated or described herein, and the objects distinguished by "first", "second", etc. are generally of one type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and / or" in the specification and claims represents at least one of the connected objects, and the character " / " generally indicates that the objects associated with each other are in an "or" relationship.

[0020] The terms "at least one (item)", "at least one of" and the like in the specification and claims of the present application refer to any one, any two or a combination of more than two of the objects included therein. For example, at least one (item) of a, b, and c can be represented by: "a", "b", "c", "a and b", "a and c", "b and c" and "a, b and c", where a, b, and c can be single or multiple. Similarly, "at least two (items)" refers to two or more, and its meaning is similar to that of "at least one (item)".

[0021] The conference task processing method provided in the embodiment of the present application can be applied to the conference recording scenario.

[0022] During a meeting, an electronic device records the speech audio of a speaker of the meeting, converts the audio into a text-based meeting record, extracts at least one meeting task information from the record, and then generates a task map of the meeting based on the at least one meeting task information. The task map is used to represent the at least one meeting task information in a graph structured manner in the form of task nodes and edges, wherein a task node represents the task content, task handler or task deadline information in a meeting task information, and the edges (i.e., connections) between multiple task nodes represent the association between the multiple task nodes. For example, the task content of a task information of the meeting is: communicating with customers, the handler of the task information is: Xiao Ming, and the deadline of the task information is a press conference next month. The generated task map representing the task information includes three task nodes, which are: "Handler: Xiao Ming", "Task: Communicating with Customers" and "Deadline: Press Conference Next Month". In this way, the task content, task handler and task deadline of the meeting task, as well as the association between the various contents are intuitively displayed through the task map, so that users can quickly locate the meeting tasks they are concerned about by viewing the task map, thereby improving the efficiency of searching for meeting content.

[0023] The execution subject of the conference task processing method provided in the embodiment of the present application may be a conference task processing device, which may be an electronic device, or a functional module or functional entity in an electronic device. The conference task processing method provided in the embodiment of the present application is described below by taking the execution of the conference task processing method by an electronic device as an example.

[0024] The conference task processing method provided by the embodiment of the present application is described in detail below through specific embodiments and application scenarios in conjunction with the accompanying drawings.

[0025] Figure 1 A flowchart of a conference task processing method provided in an embodiment of the present application is shown in FIG. Figure 1 As shown, the conference task processing method may include the following steps 201 and 202:

[0026] Step 201: The electronic device determines at least one first meeting task information of the target meeting according to the first meeting record of the target meeting.

[0027] In some embodiments of the present application, the above-mentioned target meeting may include an offline meeting, or the target meeting may also include various forms of online network meetings, such as telephone conferences, video conferences, etc.

[0028] In some embodiments of the present application, the first meeting record of the target meeting may be a meeting record including pictures, voice, and text content. Exemplarily, the meeting record may include voice data of the target meeting, text content of the target meeting, text content displayed in the target meeting, etc. Among them, the text content of the target meeting may be text content generated based on the voice data of the target meeting, and the text content displayed in the target meeting may be, for example, a document or presentation displayed on a screen.

[0029] In some embodiments of the present application, the target conference voice data may include voice data collected by an electronic device during the conference, or the target conference voice data may include voice data received from other devices. Furthermore, the voice data collected by the electronic device may be conference voice data collected by a conference application with a recording function installed on the electronic device.

[0030] In some embodiments of the present application, the electronic device may collect voice data of the target meeting, and recognize the voice data to obtain the first meeting record of the target meeting, that is, the text content of the target meeting.

[0031] For example, taking the target meeting as an online meeting, when the target meeting is needed, the electronic device establishes an online conference room through a conference application, and adds the electronic devices of other participants to the online conference room. During the meeting, the electronic device collects the speaker's voice data through the conference application to obtain conference voice data, and then recognizes the conference voice data to obtain the first meeting record.

[0032] In some embodiments of the present application, the at least one first conference task information is used to indicate at least one first conference task, and one first conference task information may indicate one first conference task.

[0033] In some embodiments of the present application, the first meeting task may be a work task assigned in the target meeting. Exemplarily, the first meeting task information may include task content, task handler information, and task deadline information.

[0034] It can be understood that the task content is the task details, which are used to describe the specific tasks or work items that need to be completed; the task handler is the task owner, who refers to the designated person who is responsible for or handles the task; while the deadline indicates the expected completion date or time node of the task.

[0035] In some embodiments of the present application, the electronic device may perform semantic analysis and entity recognition on the first meeting record to obtain at least one meeting task information.

[0036] Exemplarily, the electronic device performs semantic analysis and entity recognition on the first meeting record to obtain meeting task information of meeting task 1 and meeting task information of meeting task 2, wherein the meeting task information of meeting task 1 includes: task content: complete market research report, meeting handler: Li Hua, meeting deadline: next Friday; the meeting task information of meeting task 2 includes: task content: communicate with customers, task handler: Zhang Li, task deadline: press conference next month.

[0037] Step 202: The electronic device generates a task map of the target meeting based on at least one first meeting task information.

[0038] The task graph is used to represent at least one first meeting task information in a graph structured manner in the form of task nodes and edges. A first meeting task information includes at least one of the following: task content, task handler information, and task deadline information.

[0039] In some embodiments of the present application, the above-mentioned task map includes at least one node (Vertex) and at least one edge (Edge). Among them, a node represents an entity, that is, it represents the task content, task handler information or task deadline information; the edge represents the relationship between entities, for example, the edge formed by the task content node and the task handler information node represents that there is an association relationship between the task content and the task handler, that is, the task content is the responsibility of the task handler.

[0040] It should be noted that the above-mentioned nodes can be called task nodes, and the above-mentioned task graph can also be called knowledge graph.

[0041] In some embodiments of the present application, the electronic device may infer at least one meeting task information through a graph neural network (GNN) and generate a task graph based on the inference results.

[0042] In some embodiments of the present application, the above-mentioned reasoning results may include a classification result and a regression value of the node, the classification result is used to predict the category label of the node, and the regression value is used to predict the continuous value or numerical attribute of the node.

[0043] It should be noted that GNN is a deep learning method based on graph structure, which uses the connection relationship between nodes and graph convolution mechanism to carry out intelligent decision support and knowledge graph construction, so as to make intelligent decisions in task allocation and priority adjustment. In GNN, each word or phrase in the text is regarded as a node. The representation of the node usually comes from the pre-trained vector matrix, which will be updated during the training process to better adapt to the specific task. GNN updates the representation of the node by aggregating the information of neighboring nodes and uses different loss functions to optimize the model parameters according to the task requirements; the edge is used to represent the relationship between nodes. In the text context, this relationship can be based on co-occurrence, syntactic dependency or semantic similarity between words.

[0044] For example, at least one first task information includes the meeting task information of meeting task 1 and the meeting task information of meeting task 2, where the meeting task information of meeting task 1 includes: task content: complete market research report, meeting handler: Li Hua, meeting deadline: next Friday; the meeting task information of meeting task 2 includes: task content: communicate with customers, task handler: Zhang Li, task deadline: press conference next month, the electronic device can infer the above meeting task information through the graph neural network to generate the following Figure 2 The task map shown in Figure 2 As shown, the task map includes six task nodes, which respectively represent the task handler "Zhang Li", the task content of Task 2 "Communicate with customers", the task deadline "Press conference next month", the task handler "Li Hua", the task content of Task 1 "Complete the market research report" and the task deadline "Next Friday". Among them, the arrow between the task handler "Zhang Li" and the task content "Communicate with customers" node represents an edge, indicating that Task 2 is the responsibility of the task handler "Zhang Li", and the arrow between the task content of Task 2 and the task deadline "Press conference next month" node represents an edge, indicating that the deadline of Task 2 is the press conference next month.

[0045] In related technologies, a "one-size-fits-all" approach is often used in information display, lacking personalized customization. This model leads to inefficiency in users' search for discussion content related to themselves, increasing information redundancy and users' time costs. In addition, existing systems are unable to customize views based on personalized information such as user roles, topics of concern, and task responsibilities, affecting users' sense of participation and information acquisition efficiency.

[0046] In an embodiment of the present application, the lines (edges) between nodes in the task map represent the relationship between tasks, responsible persons, and deadlines. Through such a visualization method, users can easily view the progress status of the current task, identify the dependencies between key people and tasks, and ensure that the project is carried out on time and efficiently. This map generation method can be applied to various meeting management and task allocation scenarios, helping team members to better collaborate and communicate, and providing decision support for managers. Through real-time updates and intelligent analysis, the system can continuously optimize task allocation and resource management, and improve the team's work efficiency and project success rate.

[0047] The conference task processing method provided in the embodiment of the present application is that the electronic device determines at least one first conference task information of the target conference based on the first conference record of the target conference, and generates a task map of the target conference based on the at least one first conference task information, wherein the above-mentioned task map is used to represent the at least one first conference task information in a graph structured manner in the form of task nodes and edges, and one conference task information includes at least one of the following: task content, task handler information, and conference task deadline. Through this method, the electronic device can extract specific conference task information based on the conference record, and generate a recall task map in the form of task nodes and edges, so that the user can intuitively view the various tasks in the meeting through the task map, as well as the relationship between the task content, task handler, and task deadline of each task. In this way, the process of viewing the content of the meeting subsequently is simplified, so that the user can quickly locate the conference tasks that he is concerned about, thereby improving the efficiency of searching the content of the meeting.

[0048] In some embodiments of the present application, the above 202 may include the following steps 202a and 202b:

[0049] Step 202a: The electronic device inputs a vector matrix, an adjacency matrix, and a feature matrix corresponding to at least one first conference task information into a graph neural network model.

[0050] Step 202b: The electronic device uses a graph neural network to infer at least one first meeting task information based on a vector matrix, an adjacency matrix, and a feature matrix, and generates a task graph for the target meeting based on the inference result.

[0051] Among them, the above-mentioned vector matrix is ​​used to represent the text feature information of at least one meeting task information, the above-mentioned adjacency matrix is ​​used to represent the association between the task content, task handler information and task deadline information in at least one meeting task information, and the above-mentioned feature matrix is ​​used to represent the urgency and complexity of the meeting task corresponding to at least one meeting task information.

[0052] In some embodiments of the present application, the above-mentioned vector matrix may be composed of multiple feature vectors, and the vector matrix may represent key information and features in the text.

[0053] In some embodiments of the present application, the electronic device can select a feature extraction method to perform feature extraction on at least one meeting task information based on the characteristics of the text data and the task requirements, so as to convert the text of each meeting task information into a feature vector that represents the key information and features in the text, and then arrange the feature vectors of each meeting task information in a certain order to form a two-dimensional vector matrix, in which each row of the vector matrix can represent a feature vector of a meeting task information, and each column can represent a feature dimension.

[0054] In some embodiments of the present application, the adjacency matrix may be a vector matrix of M rows and N columns, where the rows and columns of the matrix represent different nodes, and the element A[i,j] in the matrix represents whether the node in the i-th row is connected to the node in the j-th column. 1 indicates that there is a connection, and 0 indicates that there is no connection.

[0055] For ease of description, the above adjacency matrix can be expressed as matrix A.

[0056] For example, the adjacency matrix A represents the connection relationship between the task content, task handler, and task deadline nodes of the meeting task. Through the adjacency matrix, the electronic device can generate and update the dependency structure between tasks, as well as the association between task handlers, task content, and task deadline.

[0057] In some embodiments of the present application, the feature matrix can be a vector matrix with K rows and L columns. The feature matrix is ​​used to provide attribute information of each node, such as the urgency and complexity of the meeting task. The above information is input into the GNN, so that the task allocation and decision-making process can be optimized in the dynamically changing meeting scene, and the characteristics of the task node directly affect the system's evaluation of the task priority and the adjustment of the allocation strategy.

[0058] Exemplarily, the feature matrix X is composed of an urgency coefficient and a complexity coefficient of at least one conference task, wherein a row is composed of an urgency coefficient and a complexity coefficient of a conference task.

[0059] In an embodiment of the present application, the adjacency matrix helps to transfer and associate information between different nodes; the feature matrix enriches the characteristics of each node, so that more intelligent task allocation and decision-making can be made. Specifically, through the adjacency matrix, tasks can establish deeper connections with relevant handlers, similar tasks, and deadlines, which means that more valuable information can be learned from the allocation and execution results of other tasks, and then more optimized decisions can be made. The feature matrix not only records the urgency and complexity of the task, but also can be combined with the status and other contextual information of the person in charge. This information is updated through multiple convolutions in the GNN, which can help the system more accurately evaluate the rationality of task priorities and the allocation of responsible persons. By introducing the adjacency matrix and the feature matrix, the relationship between these tasks, handlers, and deadlines is clearly represented in the form of a graph.

[0060] In some embodiments of the present application, GNN can aggregate the features of each node with the features of its surrounding nodes to form a new node representation. This process can be implemented through message passing, where each node receives messages from its neighboring nodes and aggregates these messages into a new node representation. This method can be iterated multiple times to obtain more comprehensive graph structure information.

[0061] In an embodiment of the present application, the at least one first meeting task information is inferred based on the vector matrix, adjacency matrix and feature matrix corresponding to the task information through graph neural network (GNN) technology to generate a task graph. Through the adjacency matrix and the feature matrix, the association between task nodes, handler nodes and deadline nodes can be dynamically captured, and the urgency and complexity of the task can be combined to generate a personalized task graph, thereby enhancing the efficiency of task allocation and decision execution, and ultimately improving the intelligence level and overall effect of meeting management.

[0062] In some embodiments of the present application, before the above step 202a, the conference task processing method provided by the embodiment of the present application may further include the following steps 202c and 202d:

[0063] Step 202c: The electronic device generates at least one adjacency matrix according to the association relationship among the task content, task handler information, and task deadline information of at least one first conference task information.

[0064] Step 202d: The electronic device generates at least one feature matrix according to the urgency coefficient and complexity coefficient of the conference task corresponding to the at least one first conference task information.

[0065] Among them, the above-mentioned urgency coefficient is determined based on the deadline information of the corresponding meeting task, and the above-mentioned urgency coefficient is used to characterize the urgency of the corresponding meeting task; the above-mentioned complexity coefficient is determined based on the number of task nodes associated with the task content of the corresponding meeting task, and the above-mentioned complexity coefficient is used to characterize the complexity of the corresponding meeting task.

[0066] In some embodiments of the present application, the electronic device can analyze the content of each meeting task, determine the relationship between the meeting tasks, and generate a two-dimensional matrix based on the relationship between the task content, task handlers, and task deadlines of each meeting task and the relationship between the meeting tasks.

[0067] For example, at least one meeting task information includes meeting information of meeting task 1 and meeting task 2, where the meeting task information of meeting task 1 includes: task content: complete market research report, meeting handler: Li Hua, meeting deadline: next Friday; the meeting task information of meeting task 2 includes: task content: communicate with customers, task handler: Zhang Li, task deadline: press conference next month, and the adjacency matrix A generated according to the meeting task information of meeting task 1 and the meeting task information of meeting task 2 is as follows:

[0068]

[0069] The labels of the rows and columns of the adjacency matrix A are nodes, including tasks T1 (market research report), T2 (communication with customers), responsible person Li Hua, responsible person Zhang Li, deadline nodes "next Friday" and "press conference". For example, matrix A[1,3]=1 indicates that there is a connection between task T1 and Li Hua, indicating that Li Hua is responsible for task T1. Similarly, A[2,6]=1 indicates that there is a connection between task T2 and the press conference, indicating that the deadline of task T2 is the press conference.

[0070] In some embodiments of the present application, the electronic device can determine the urgency coefficient of at least one conference task based on the deadline information of the conference task corresponding to at least one conference task information, and determine the complexity coefficient of at least one conference task based on the number of task nodes associated with the at least one conference task, and then generate a two-dimensional vector matrix based on the urgency coefficient and complexity coefficient of the at least one conference task.

[0071] For example, the urgency coefficient represents the urgency or urgency of the meeting task. The urgency coefficient U i It can be calculated by the following formula (1):

[0072]

[0073] Among them, the remaining time is the number of days from the deadline of the meeting task to the current day. The fewer the days, the higher the urgency. α is the importance coefficient of the task, which is used to adjust the urgency of the task.

[0074] It should be noted that α can be set according to the importance or priority of the task.

[0075] For example, taking at least one first meeting task information including the meeting task information of meeting task 1 and the meeting task information of meeting task 2 as an example, assuming that the deadline of meeting task 1 is Friday, the current date is Monday, the remaining time is 4 days, and α=0.9, then the urgency coefficient of meeting task 1 can be calculated as 0.225 according to the remaining time through the above formula (1), that is, Assuming that the remaining time of meeting task 2 is 30 days and α = 0.7, the urgency coefficient of meeting task 2 can be calculated as 0.023 according to the remaining time using the above formula (1), that is,

[0076] It can be understood that the urgency coefficient of conference task 1 is greater than the urgency coefficient of conference task 2 , indicating that the urgency of conference task 1 is higher than that of conference task 2 .

[0077] For example, the complexity coefficient represents the complexity or complexity of the meeting task. i It can be calculated by the following formula (2):

[0078]

[0079] Among them, the node degree indicates the number of other nodes connected to the task node. The larger the degree, the more connections the task has with other nodes, and the more complex the task may be; the shortest path length indicates the shortest path length from the task node to its deadline node. The shorter the path, the easier it is to plan and complete the task, and the lower the complexity. This can be done by finding the shortest path from the task node to the deadline node in the adjacency matrix. For example, the path from the task to "next Friday" has only one direct edge, so the path length is 1; the clustering coefficient indicates the degree of interconnection between the task node and its neighboring nodes. That is, if the nodes connected to a task also have many connections with each other, it means that the task is in a high-density area in the network, and the complexity will increase. The clustering coefficient can be calculated by the following formula:

[0080] Clustering coefficient = actual number of connected edges / maximum possible number of connected edges

[0081] For example, for the above meeting task 1 (market research report), the node degree is 2 (connecting Li Hua and next Friday), the shortest path length is 1 (an edge from T1 to the deadline node "next Friday"), and the clustering coefficient is 1 (there is an edge between Li Hua and next Friday, and the maximum possible number of connections is also 1, so the clustering coefficient is 1). Assuming that the weight coefficients λ_1=0.3, λ_2=0.4, λ_3=0.3, the complexity coefficient of the meeting task 1 can be calculated by the above formula (2) as 1.3, that is, C T1 =0.3×2+0.4×1+0.3×1=1.3. Similarly, for the above meeting task T2 (communication with customers), it can be calculated that the complexity coefficient of the meeting task 1 is 1.3, C T2 =0.3×2+0.4×0.5+0.3×1=1.1.

[0082] Exemplarily, taking at least one conference task information including conference task 1 and conference task 2 as an example, the urgency coefficient of conference task 1 (T1) is: 0.225, the complexity coefficient of conference task 1 is 1.3, the urgency coefficient of conference task 2 (T2) is: 0.023, the complexity coefficient of conference task 2 is 1.1, the special matrix X generated according to the urgency coefficient and complexity coefficient of conference task 1, and the urgency coefficient and complexity coefficient of conference task 2 is as follows:

[0083]

[0084] Exemplarily, the feature matrix X represents the features of each node, including urgency and complexity. The rows of the feature matrix represent each node (task, handler, deadline, etc.), and the columns represent the features of the node (urgency, complexity).

[0085] In some embodiments of the present application, after the above step 202, the conference task processing method provided by the embodiment of the present application may further include the following steps 203 to 205:

[0086] Step 203: The electronic device determines at least one second meeting task information of the target meeting according to the second meeting record of the target meeting.

[0087] Step 204: The electronic device inputs a vector matrix, an adjacency matrix, and a feature matrix corresponding to at least one second conference task information into the graph neural network model.

[0088] Step 205: The electronic device uses a graph neural network to infer at least one second meeting task information based on the vector matrix, adjacency matrix and feature matrix corresponding to the second meeting task information, and updates the task graph according to the inference result.

[0089] It should be noted that the explanation of the vector matrix, adjacency matrix and feature matrix corresponding to at least one second conference task information can be found in the relevant description of the vector matrix, adjacency matrix and feature matrix corresponding to at least one second conference task information in the above embodiments, which will not be repeated here.

[0090] In some embodiments of the present application, the electronic device may obtain at least one second meeting task information based on the results of recognition and semantic analysis of the new meeting record, and update the task map in real time based on the at least one second meeting task information.

[0091] Exemplarily, when the electronic device obtains the new meeting record obtained by real-time speech conversion, it also performs semantic understanding and entity recognition on the meeting record to obtain at least one meeting task information of the meeting record, and then inputs the vector matrix, adjacency matrix and feature matrix of at least one meeting task information into the GNN. Based on the vector matrix, adjacency matrix and feature matrix of at least one meeting task information, the GNN performs model reasoning on the at least one meeting task information to obtain the reasoning result, and updates the task map according to the reasoning result. In this cycle, when the meeting is ongoing, the task map will be continuously updated to obtain all tasks and matters involved in different subjects in the meeting.

[0092] For example, it is assumed that the following meeting tasks are identified in the target meeting:

[0093] Task 1: Task content: Complete the market research report, Task handler: Li Hua, Task deadline: Next Friday;

[0094] Task 2: Task content: communicate with customers, task handler: Zhang Li, task deadline: next month's press conference;

[0095] Task 3: Task content: prepare materials for new product launch conference, task handler: Li Hua, task deadline: next Wednesday;

[0096] Task 4: Task content: Analyze competitor reports, Task handler: Zhang Li, Task deadline: two weeks later;

[0097] Task 5: Task content: Develop a quarterly marketing plan, Task handlers: Li Hua and Zhang Li, Task deadline: three weeks later;

[0098] After updating the above task map according to the above meeting information, the updated task map is as follows Figure 3 shown.

[0099] In some embodiments of the present application, after the electronic device inputs the vector matrix, adjacency matrix and feature matrix corresponding to at least one second meeting task information into the graph neural network model, the graph neural network model obtains a new adjacency matrix A1 based on the adjacency matrix corresponding to at least one second meeting task information and the adjacency matrix corresponding to at least one first meeting task information, and the graph neural network model performs reasoning based on the new adjacency matrix A1 to update the task graph.

[0100] Exemplarily, assuming that the first task information includes the meeting task information of meeting task 1 and meeting task 2, and the second meeting task information includes the meeting task information of meeting task 3 and meeting task 4, the new adjacency matrix A1 is as follows:

[0101]

[0102] It should be noted that, for the convenience of description, conference task 1 can be called T1, conference task 2 can be called T2, conference task 3 can be called T3, and conference task 4 can be called T4.

[0103] It should be noted that the tasks, responsible persons and deadlines in the meeting are not static. The priority and deadline of the task may change over time, or new tasks may be added. In an embodiment of the present application, this dynamically evolving relationship can be flexibly handled through a graph neural network. Specifically, after a new task is added, the task map can be updated immediately, the processing allocation can be readjusted, or the potential relationship between the task and other tasks can be inferred (such as whether the same person is responsible). Changes in the deadline will also affect other related tasks through multiple information transfers in the task map, and the system can automatically adjust the priority and allocation strategy of the task.

[0104] In some embodiments of the present application, the electronic device may display the generated task map, thereby realizing the generation and display of a personalized task view.

[0105] Exemplarily, the task map includes the task entity, deadline, and task handler of each person in the current meeting.

[0106] Furthermore, the electronic device can extract all branches related to the user from the meeting records, sort and filter the information according to task priority, deadline, etc., and generate a personalized description for each participant.

[0107] For example: Li Hua’s task:

[0108] Task 1: Complete the market research report, deadline: next Friday; Task 3: Prepare materials for the new product launch, deadline: next Wednesday; Task 5: Develop a quarterly marketing plan, deadline: three weeks later.

[0109] Zhang Li's mission:

[0110] Task 2: Communicate with customers, deadline: next month's press conference; Task 4: Analyze competitor reports, deadline: two weeks later; Task 5: Develop a quarterly marketing plan, deadline: three weeks later.

[0111] In this way, through each person's node in the graph, all the nodes associated with it are found and diverged outward, so that everyone's meeting tasks can be clearly recorded.

[0112] In some embodiments of the present application, the above step 201 can be implemented by the following steps 201a and 201b.

[0113] Step 201a: The electronic device inputs the first meeting record of the target meeting into the deep learning model.

[0114] Step 201b: The electronic device performs semantic understanding and entity recognition on the first meeting record through a deep learning model to obtain at least one first meeting task information of the target meeting.

[0115] In some embodiments of the present application, the electronic device may perform semantic analysis on the transcribed text through a deep learning model (such as BERT or Transformer) to obtain at least one meeting task information.

[0116] It should be noted that the deep learning model can understand the context of the text and identify the sentence structure and semantic dependencies in the text. Through semantic analysis of the text, important information such as key topics, action items, and responsible persons in the speech can be identified. The entity recognition algorithm is integrated into the semantic analysis process to automatically identify and classify key entities in the text. These entities include names of people, organization names, dates, tasks, locations, etc. By combining semantic understanding and entity recognition, meeting task information in meeting minutes can be accurately extracted.

[0117] For example, the minutes of the target meeting are: "We need to complete the market research report by next Friday, and this task will be undertaken by Li Hua. In addition, Zhang Li will be responsible for communicating with customers to ensure that all requirements are met at the launch conference next month." After the deep learning model performs semantic understanding and entity recognition on the above meeting minutes, at least one meeting task information obtained includes: Task 1: Market research report, deadline: next Friday, responsible person: Li Hua, and Task 2: Communicate with customers, deadline: next Wednesday, responsible person: Zhang Li.

[0118] It should be noted that traditional meeting record systems rely on basic natural language processing (NLP) technology for semantic understanding, and are usually unable to deeply understand the complex semantic relationships of speech content. Due to the lack of dynamic context modeling capabilities, the system cannot effectively identify important information in the speech content, such as task content, task handlers, and task deadlines. This makes meeting minutes often lack structure and pertinence, making it difficult to provide effective information for subsequent task allocation and decision support.

[0119] In the embodiment of the present application, by combining semantic understanding and entity recognition, the complex semantic relationship of the speech content can be deeply understood, so as to accurately identify important information such as task content, task handlers, and task deadlines, so as to accurately generate a task map.

[0120] In some embodiments of the present application, the electronic device may convert the audio of the target meeting into text through automatic speech recognition technology (Automatic Speech Recognition, ASR), transcribe the speaker's voice content in real time, and obtain the first meeting record.

[0121] For example, the electronic device can identify the voice features of different speakers through timbre recognition, thereby distinguishing different speakers and the speech content of different speakers.

[0122] It should be noted that different people have different timbre, and the timbre of the speaker during the meeting is a relatively stable output. We can first classify the different participants through voice classification. At the same time as voice classification, the speaker's voice is transcribed, that is, as long as it can correspond to speaker A and the corresponding content of speaker A.

[0123] In the embodiments of the present application, by combining automatic speech recognition (ASR) technology and timbre classification, the voice features of different speakers are accurately identified, and combined with the semantic information of the speech content, the speaker's identity is accurately confirmed, solving the problem of high speaker recognition error rate in traditional systems.

[0124] In some embodiments of the present application, after the above step 202, the conference task processing method provided by the embodiment of the present application may further include the following steps 206 to 208:

[0125] Step 206: The electronic device generates at least one meeting schedule information based on at least one meeting task information.

[0126] Step 207: The electronic device determines at least one task handler corresponding to at least one meeting schedule information.

[0127] Among them, one meeting schedule information corresponds to one or more task handlers.

[0128] Step 208: The electronic device sends at least one meeting schedule information to a corresponding task handler.

[0129] In some embodiments of the present application, the meeting task schedule information includes task content, task handler information, and task deadline information. By creating meeting schedule information, the meeting organizer can have a comprehensive and clear task plan, thereby effectively arranging and managing the meeting, so that the task handler can clearly know his or her task arrangement.

[0130] In some embodiments of the present application, the electronic device can send the generated meeting schedule information to the corresponding task handler. For example, the electronic device can usually send the meeting schedule information through email, text message, in-app notification or other instant messaging tools. In this way, by sending the meeting schedule information to the task handler, the electronic device ensures that each relevant person can obtain the meeting plan and task requirements related to them in a timely manner, so that the task handler can prepare in advance, clarify his role and tasks in the meeting, and thus participate in and execute the meeting plan more effectively.

[0131] Exemplarily, taking at least one meeting task information including: Task 1: Market research report, deadline: next Friday, handler: Li Hua as an example, the electronic device can generate a meeting task schedule information based on the above meeting task information: "The market research report needs to be completed before Friday next week" and send the meeting task schedule information to the handler Li Hua via email.

[0132] In an embodiment of the present application, the electronic device can generate meeting schedule information based on the task map and integrate it with the schedules of the participants to realize the automation of task and schedule management, thereby ensuring the effective execution of meeting decisions.

[0133] In combination with the above embodiments, the conference task processing method provided in the embodiments of the present application is exemplarily described below by way of examples.

[0134] Exemplarily, the above conference task processing method may include the following steps:

[0135] Step 11: Speaker identification and content transcription to obtain the first meeting record.

[0136] Step 12: Perform semantic understanding and entity recognition on the first meeting record to obtain at least one first meeting task information.

[0137] Step 13: Generate a task map based on at least one first meeting task information.

[0138] Step 14: Speaker identification and content transcription to obtain the second meeting record.

[0139] Step 15: Perform semantic understanding and entity recognition on the first meeting record to obtain at least one second meeting task information.

[0140] Step 16: Based on at least one second meeting task information, update the task map.

[0141] Step 17: Display the personalized view based on the updated task map.

[0142] Step 18: Generate task schedule information based on the updated task map.

[0143] In the embodiment of this application, the conference management process is comprehensively optimized by improving the accuracy of speaker identification, enhancing semantic understanding and context modeling capabilities, introducing intelligent decision support, realizing personalized content presentation and automated task management. The system not only improves the accuracy and completeness of meeting records, but also significantly enhances the efficiency of task allocation and decision execution, ultimately improving the intelligent level and overall effect of conference management.

[0144] The above-mentioned method embodiments, or various possible implementation methods in each method embodiment, can be executed separately, or, under the premise that there is no contradiction, can also be executed in combination with each other. The specific implementation can be determined according to actual usage requirements, and the embodiments of the present application do not limit this.

[0145] The conference task processing method provided in the embodiment of the present application can be executed by a conference task processing device. The conference task processing device provided in the embodiment of the present application is described by taking the conference task processing method executed by the conference task processing device as an example.

[0146] Figure 4 A schematic diagram of the structure of the conference task processing device 400 provided in the embodiment of the present application is shown in FIG. Figure 4 As shown, the conference task device 400 may include: a processing module 401, the processing module 401 is used to: determine at least one first conference task information of the target meeting based on the first meeting record of the target meeting; generate a first task graph of the target meeting based on the at least one first meeting task information; wherein the first task graph is used to represent at least one first meeting task information in a graph structured manner in the form of task nodes and edges, and a first meeting task information includes at least one of the following: task content, task handler information, and task deadline information.

[0147] In some embodiments of the present application, the processing module is specifically used to: input the vector matrix, adjacency matrix and feature matrix corresponding to at least one first meeting task information into a graph neural network model; through the graph neural network, infer the at least one first meeting task information according to the vector matrix, adjacency matrix and feature matrix, and generate a first task graph of the target meeting according to the inference result; wherein the vector matrix is ​​used to represent the text feature information of at least one meeting task information, the adjacency matrix is ​​used to represent the association between the task content, task handler information and task deadline information in at least one meeting task information, and the feature matrix is ​​used to represent the urgency and complexity of the meeting task corresponding to at least one meeting task information.

[0148] In some embodiments of the present application, the processing module is also used to generate an adjacency matrix based on the correlation between the task content, task handler information and task deadline information of at least one first meeting task information before inputting the vector matrix, adjacency matrix and feature matrix corresponding to at least one first meeting task information into the graph neural network model; the processing module is also used to generate a feature matrix based on the urgency coefficient and complexity coefficient of the meeting task corresponding to at least one first meeting task information; wherein the urgency coefficient is determined based on the deadline information of the corresponding meeting task, and the urgency coefficient is used to characterize the urgency of the corresponding meeting task; the complexity coefficient is determined based on the number of task nodes associated with the task content of the corresponding meeting task, and the complexity coefficient is used to characterize the complexity of the corresponding meeting task.

[0149] In some embodiments of the present application, the processing module is also used to determine at least one second meeting task information of the target meeting based on the second meeting record of the target meeting after generating a first task graph of the target meeting based on at least one first meeting task information; the processing module is also used to input the vector matrix, adjacency matrix and feature matrix corresponding to the at least one second meeting task information into the graph neural network model; the processing module is also used to infer at least one second meeting task information through the graph neural network based on the vector matrix, adjacency matrix and feature matrix corresponding to the second meeting task information, and update the first task graph based on the inference result.

[0150] In some embodiments of the present application, the processing module is specifically used to: input the first meeting record of the target meeting into a deep learning model; perform semantic understanding and entity recognition on the first meeting record through the deep learning model to obtain at least one first meeting task information of the target meeting.

[0151] In some embodiments of the present application, the processing module is also used to generate at least one meeting schedule information based on at least one meeting task information after generating a first task map of the target meeting according to at least one first meeting task information; the processing module is also used to determine at least one task handler corresponding to at least one meeting schedule information, and one meeting schedule information corresponds to one or more task handlers; the sending module is used to send at least one meeting schedule information to the corresponding task handler.

[0152] The conference task processing device provided in the embodiment of the present application determines at least one first conference task information of the target conference based on the first conference record of the target conference, and generates a task map of the target conference based on the at least one first conference task information, wherein the above-mentioned task map is used to represent at least one first conference task information in a graph structured manner in the form of task nodes and edges, and one first conference task information includes at least one of the following: task content, task handler information, and task deadline information. Through this method, the electronic device can extract specific conference task information based on the conference record, and generate a task map in the form of task nodes and edges, so that the user can intuitively view the various tasks in the meeting through the task map, as well as the relationship between the task content, task handler, and task deadline of each task. In this way, the subsequent process of viewing the content of the meeting is simplified, allowing users to quickly locate the conference tasks they are concerned about, thereby improving the efficiency of searching the content of the meeting.

[0153] The conference task processing device in the embodiment of the present application can be an electronic device, or a component in the electronic device, such as an integrated circuit or a chip. The electronic device can be a terminal, or it can be other devices other than a terminal. Exemplarily, the electronic device can be a mobile phone, a tablet computer, a laptop computer, a PDA, a vehicle-mounted electronic device, a mobile Internet device (Mobile Internet Device, MID), an augmented reality (augmented reality, AR) / virtual reality (virtual reality, VR) device, a robot, a wearable device, an ultra-mobile personal computer (ultra-mobile personal computer, UMPC), a netbook or a personal digital assistant (personal digital assistant, PDA), etc., and can also be a server, a network attached storage (Network Attached Storage, NAS), a personal computer (personal computer, PC), a television (television, TV), a teller machine or a self-service machine, etc., which is not specifically limited in the embodiment of the present application.

[0154] The conference task processing device in the embodiment of the present application may be a device having an operating system. The operating system may be an Android operating system, an iOS operating system, or other possible operating systems, which are not specifically limited in the embodiment of the present application.

[0155] The conference task processing device provided in the embodiment of the present application can implement each process implemented by the method embodiment of the conference task processing method. To avoid repetition, it will not be described here.

[0156] Alternatively, if Figure 5 As shown, an embodiment of the present application also provides an electronic device 500, including a processor 501 and a memory 502, wherein the memory 502 stores programs or instructions that can be executed on the processor 501, and when the program or instructions are executed by the processor 501, the various steps of the above-mentioned conference task processing method embodiment are implemented, and the same technical effect can be achieved. To avoid repetition, they are not described here.

[0157] It should be noted that the electronic devices in the embodiments of the present application include the mobile electronic devices and non-mobile electronic devices mentioned above.

[0158] Figure 6 A schematic diagram of the hardware structure of an electronic device to implement an embodiment of the present application.

[0159] The electronic device 100 includes but is not limited to components such as a radio frequency unit 101, a network module 102, an audio output unit 103, an input unit 104, a sensor 105, a display unit 106, a user input unit 107, an interface unit 108, a memory 109, and a processor 110.

[0160] Those skilled in the art will appreciate that the electronic device 100 may also include a power source (such as a battery) for supplying power to various components, and the power source may be logically connected to the processor 110 through a power management system, thereby implementing functions such as managing charging, discharging, and power consumption management through the power management system. Figure 6 The electronic device structure shown in the figure does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently, which will not be described in detail here.

[0161] Among them, the processor 110 is used to: determine at least one first meeting task information of the target meeting based on the first meeting record of the target meeting; generate a first task graph of the target meeting based on the at least one first meeting task information; wherein the first task graph is used to represent at least one first meeting task information in a graph structured manner in the form of task nodes and edges, and a first meeting task information includes at least one of the following: task content, task handler information, and task deadline information.

[0162] In some embodiments of the present application, the processor 110 is specifically used to: input a vector matrix, an adjacency matrix and a feature matrix corresponding to at least one first meeting task information into a graph neural network model; through the graph neural network, infer at least one first meeting task information according to the vector matrix, the adjacency matrix and the feature matrix, and generate a first task graph of the target meeting according to the inference result; wherein the vector matrix is ​​used to represent the text feature information of at least one meeting task information, the adjacency matrix is ​​used to represent the association between the task content, the task handler information and the task deadline information in at least one meeting task information, and the feature matrix is ​​used to represent the urgency and complexity of the meeting task corresponding to at least one meeting task information.

[0163] In some embodiments of the present application, the processor 110 is also used to generate an adjacency matrix based on the correlation between the task content, task handler information and task deadline information of at least one first meeting task information before inputting the vector matrix, adjacency matrix and feature matrix corresponding to at least one first meeting task information into the graph neural network model; the processor 110 is also used to generate a feature matrix based on the urgency coefficient and complexity coefficient of the meeting task corresponding to at least one first meeting task information; wherein the urgency coefficient is determined based on the deadline information of the corresponding meeting task, and the urgency coefficient is used to characterize the urgency of the corresponding meeting task; the complexity coefficient is determined based on the number of task nodes associated with the task content of the corresponding meeting task, and the complexity coefficient is used to characterize the complexity of the corresponding meeting task.

[0164] In some embodiments of the present application, the processor 110 is also used to determine at least one second meeting task information of the target meeting based on the second meeting record of the target meeting after generating a first task graph of the target meeting based on at least one first meeting task information; the processor 110 is also used to input the vector matrix, adjacency matrix and feature matrix corresponding to the at least one second meeting task information into the graph neural network model; the processor 110 is also used to infer at least one second meeting task information through the graph neural network based on the vector matrix, adjacency matrix and feature matrix corresponding to the second meeting task information, and update the first task graph based on the inference result.

[0165] In some embodiments of the present application, the processor 110 is specifically used to: input the first meeting record of the target meeting into a deep learning model; perform semantic understanding and entity recognition on the first meeting record through the deep learning model to obtain at least one first meeting task information of the target meeting.

[0166] In some embodiments of the present application, the processor 110 is also used to generate at least one meeting schedule information based on at least one meeting task information after generating a first task map of the target meeting according to at least one first meeting task information; the processor 110 is also used to determine at least one task handler corresponding to at least one meeting schedule information, and one meeting schedule information corresponds to one or more task handlers; the radio frequency unit 101 is used to send at least one meeting schedule information to the corresponding task handler.

[0167] The electronic device provided in the embodiment of the present application determines at least one first meeting task information of the target meeting based on the first meeting record of the target meeting, and generates a task map of the target meeting based on the at least one first meeting task information, wherein the above-mentioned task map is used to represent the at least one first meeting task information in a graph structured manner in the form of task nodes and edges, and a first meeting task information includes at least one of the following: task content, task handler information, and task deadline information. Through this method, the electronic device can extract specific meeting task information based on the meeting record, and generate a task map in the form of task nodes and edges, so that the user can intuitively view the various tasks in the meeting through the task map, as well as the relationship between the task content, task handler, and task deadline of each task. In this way, the process of viewing the content of the meeting subsequently is simplified, so that the user can quickly locate the meeting tasks that he is concerned about, thereby improving the efficiency of searching the content of the meeting.

[0168] It should be understood that in the embodiment of the present application, the input unit 104 may include a graphics processor (Graphics Processing Unit, GPU) 1041 and a microphone 1042, and the graphics processor 1041 processes the image data of a static picture or video obtained by an image capture device (such as a camera) in a video capture mode or an image capture mode. The display unit 106 may include a display panel 1061, and the display panel 1061 may be configured in the form of a liquid crystal display, an organic light emitting diode, etc. The user input unit 107 includes a touch panel 1071 and at least one of other input devices 1072. The touch panel 1071 is also called a touch screen. The touch panel 1071 may include two parts: a touch detection device and a touch controller. Other input devices 1072 may include, but are not limited to, a physical keyboard, function keys (such as volume control keys, switch keys, etc.), a trackball, a mouse, and a joystick, which will not be repeated here.

[0169] The memory 109 can be used to store software programs and various data. The memory 109 may mainly include a first storage area for storing programs or instructions and a second storage area for storing data, wherein the first storage area may store an operating system, an application program or instructions required for at least one function (such as a sound playback function, an image playback function, etc.), etc. In addition, the memory 109 may include a volatile memory or a non-volatile memory, or the memory 109 may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), a static random access memory (SRAM), a dynamic random access memory (DRAM), a synchronous dynamic random access memory (SDRAM), a double data rate synchronous dynamic random access memory (DDRSDRAM), an enhanced synchronous dynamic random access memory (ESDRAM), a synchronous link dynamic random access memory (SLDRAM) and a direct memory bus random access memory (DRRAM). The memory 109 in the embodiment of the present application includes but is not limited to these and any other suitable types of memory.

[0170] The processor 110 may include one or more processing units; optionally, the processor 110 integrates an application processor and a modem processor, wherein the application processor mainly processes operations related to an operating system, a user interface, and application programs, and the modem processor mainly processes wireless communication signals, such as a baseband processor. It is understandable that the modem processor may not be integrated into the processor 110.

[0171] An embodiment of the present application also provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, each process of the above-mentioned conference task processing method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0172] The processor is the processor in the electronic device described in the above embodiment. The readable storage medium includes a computer readable storage medium, such as a computer read-only memory ROM, a random access memory RAM, a magnetic disk or an optical disk.

[0173] An embodiment of the present application further provides a chip, which includes a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the various processes of the above-mentioned conference task processing method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0174] It should be understood that the chip mentioned in the embodiments of the present application can also be called a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.

[0175] An embodiment of the present application provides a computer program product, which is stored in a storage medium. The program product is executed by at least one processor to implement the various processes of the above-mentioned conference task processing method embodiment and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0176] It should be noted that, in this article, the terms "comprise", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise one..." do not exclude the presence of other identical elements in the process, method, article or device including the element. In addition, it should be pointed out that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved, for example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.

[0177] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a computer software product, which is stored in a storage medium (such as ROM / RAM, a disk, or an optical disk), and includes a number of instructions for a terminal (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0178] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose of the present application and the scope of protection of the claims, all of which are within the protection of the present application.

Claims

1. A task processing method, characterized in that: The method comprises: Determining at least one first meeting task information of the target meeting according to the first meeting record of the target meeting; Generate a first task map of the target meeting according to the at least one first meeting task information; Among them, the first task graph is used to represent the at least one first meeting task information in a graph structured manner in the form of task nodes and edges, and the first meeting task information includes at least one of the following: task content, task handler information, and task deadline information.

2. The method according to claim 1, characterized in that: The step of generating a first task map of the target meeting according to the at least one first meeting task information includes: Inputting the vector matrix, the adjacency matrix and the feature matrix corresponding to the at least one first meeting task information into the graph neural network model; Using the graph neural network, reasoning the at least one first meeting task information according to the vector matrix, the adjacency matrix, and the feature matrix, and generating a first task graph of the target meeting according to the reasoning result; Among them, the vector matrix is ​​used to represent the text feature information of the at least one meeting task information, the adjacency matrix is ​​used to represent the association between the task content, task handler information and task deadline information in the at least one meeting task information, and the feature matrix is ​​used to represent the urgency and complexity of the meeting task corresponding to the at least one meeting task information.

3. The method according to claim 2, characterized in that Before inputting the vector matrix, the adjacency matrix and the feature matrix corresponding to the at least one first meeting task information into the graph neural network model, the method further includes: Generate an adjacency matrix according to the association relationship between the task content, task handler information and task deadline information of the at least one first meeting task information; Generate a feature matrix according to the urgency coefficient and complexity coefficient of the conference task corresponding to the at least one first conference task information; Among them, the urgency coefficient is determined based on the deadline information of the corresponding meeting task, and the urgency coefficient is used to characterize the urgency of the corresponding meeting task; the complexity coefficient is determined based on the number of task nodes associated with the task content of the corresponding meeting task, and the complexity coefficient is used to characterize the complexity of the corresponding meeting task.

4. The method according to any one of claims 1 to 3, characterized in that After generating the first task map of the target meeting according to the at least one first meeting task information, the method further includes: Determining at least one second meeting task information of the target meeting according to the second meeting record of the target meeting; Inputting the vector matrix, the adjacency matrix and the feature matrix corresponding to the at least one second meeting task information into the graph neural network model; Through the graph neural network, the at least one second meeting task information is inferred according to the vector matrix, adjacency matrix and feature matrix corresponding to the second meeting task information, and the first task graph is updated according to the inference result.

5. The method according to claim 1, characterized in that After generating the first task map of the target meeting according to the at least one first meeting task information, the method further includes: Based on the at least one meeting task information, generating at least one meeting schedule information; Determine at least one task handler corresponding to the at least one conference schedule information, where one conference schedule information corresponds to one or more task handlers; The at least one meeting schedule information is sent to a corresponding task handler.

6. A conference task processing device, characterized in that: The device comprises: a processing module, wherein the processing module is used to: Determining at least one first meeting task information of the target meeting according to the first meeting record of the target meeting; Generate a first task map of the target meeting according to the at least one first meeting task information; Among them, the first task graph is used to represent the at least one first meeting task information in a graph structured manner in the form of task nodes and edges, and the first meeting task information includes at least one of the following: task content, task handler information, and task deadline information.

7. The device according to claim 6, characterized in that The processing module is specifically used for: Inputting the vector matrix, the adjacency matrix and the feature matrix corresponding to the at least one first meeting task information into the graph neural network model; Using the graph neural network, reasoning the at least one first meeting task information according to the vector matrix, the adjacency matrix, and the feature matrix, and generating a first task graph of the target meeting according to the reasoning result; Among them, the vector matrix is ​​used to represent the text feature information of the at least one meeting task information, the adjacency matrix is ​​used to represent the association between the task content, task handler information and task deadline information in the at least one meeting task information, and the feature matrix is ​​used to represent the urgency and complexity of the meeting task corresponding to the at least one meeting task information.

8. The device according to claim 7, characterized in that The processing module is further used to generate an adjacency matrix according to the association relationship between the task content, task handler information and task deadline information of the at least one first meeting task information before inputting the vector matrix, adjacency matrix and feature matrix corresponding to the at least one first meeting task information into the graph neural network model; The processing module is further used to generate a feature matrix according to the urgency coefficient and complexity coefficient of the conference task corresponding to the at least one first conference task information; Among them, the urgency coefficient is determined based on the deadline information of the corresponding meeting task, and the urgency coefficient is used to characterize the urgency of the corresponding meeting task; the complexity coefficient is determined based on the number of task nodes associated with the task content of the corresponding meeting task, and the complexity coefficient is used to characterize the complexity of the corresponding meeting task.

9. The device according to any one of claims 6 to 8, characterized in that The processing module is further configured to determine at least one second meeting task information of the target meeting according to the second meeting record of the target meeting after generating the first task map of the target meeting according to the at least one first meeting task information; The processing module is further used to input the vector matrix, adjacency matrix and feature matrix corresponding to the at least one second conference task information into the graph neural network model; The processing module is also used to infer the at least one second meeting task information through the graph neural network according to the vector matrix, adjacency matrix and feature matrix corresponding to the second meeting task information, and update the first task graph according to the inference result.

10. The device according to claim 6, characterized in that The processing module is further configured to generate at least one meeting schedule information based on the at least one meeting task information after generating the first task map of the target meeting according to the at least one first meeting task information; The processing module is further used to determine at least one task handler corresponding to the at least one conference schedule information, and one conference schedule information corresponds to one or more task handlers; The sending module is used to send the at least one conference schedule information to the corresponding task handler.

Images