Task processing method and device

Abstract

The invention provides a task processing method and device, and relates to the technical field of artificial intelligence. The task processing method comprises the following steps: generating a knowledge hypergraph by utilizing a preset large model according to original data and task processing cues; generating context parameters of the rule according to a task processing demand input by a user; according to the knowledge hypergraph and the context parameters, utilizing a preset large model to generate a plurality of task processing schemes, and displaying the task processing schemes and the knowledge hypergraph through an interactive interface; and in response to the scheme updating instruction, updating the task processing scheme. According to the technical scheme, automatic decision making efficiency and manual intervention controllability are both considered, candidate schemes can be rapidly generated through a large model, reasoning efficiency is improved, and the defects of insufficient flexibility and accuracy during pure automatic scheme generation are overcome by means of updating of user interaction participation schemes; and the accuracy and the interpretability are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, in particular to a task processing method and device. BACKGROUND

[0002] At present, since the knowledge hypergraph can represent the relationship between multiple entities and can handle high-order correlation, by using hypergraph calculation, intelligent behavior in human-machine environment systems can be better understood and modeled, and more effective algorithms and models can be developed, and even some research uses knowledge hypergraph to assist humans in decision-making.

[0003] However, in a complex decision-making process, the existing complex knowledge hypergraph is difficult to understand, cannot directly support decision-making and artificial intelligence assisted decision-making such as large models, and at the same time of poor interpretability and reasoning efficiency, it will lead to low accuracy. SUMMARY

[0004] Based on this, the present application provides a task processing method and device, which realizes higher efficiency while further improving accuracy and interpretability.

[0005] According to an aspect of the present application, a task processing method is provided, comprising: generating a knowledge hypergraph using a preset large model according to original data and task processing prompt words; generating a rule context parameter according to a user input task processing requirement; generating a plurality of task processing schemes using a preset large model according to the knowledge hypergraph and the context parameter, and displaying the task processing schemes and the knowledge hypergraph through an interactive interface; and updating the task processing schemes in response to a scheme update instruction, wherein the scheme update instruction is triggered in the case of original data update, task processing requirement modification and / or task processing scheme modification, and the task processing requirement modification and the task processing scheme modification are realized by the user based on the interactive interface.

[0006] According to some embodiments, the knowledge hypergraph includes: an entity element layer for storing attribute values and state values of entities and relationships between entities; a matter concept layer for storing logic rule data, wherein the logic rule data is used to describe triggering conditions and / or logic of events; and a control execution layer for generating an execution scheme according to the logic rule data and the state values of the entities.

[0007] According to some embodiments, the task processing prompt words include an entity description template of the entity element layer, an event description template of the matter concept layer, a behavior tree framework template of the control execution layer, and an output sample example.

[0008] According to some embodiments, a plurality of task processing schemes are generated by using a preset large model according to the knowledge hypergraph and the context parameters, and the task processing schemes and the knowledge hypergraph are displayed through an interactive interface, including: adjusting the execution scheme of the control execution layer based on the knowledge hypergraph by using the preset large model according to the context parameters, generating a plurality of task processing schemes; visualizing the knowledge hypergraph and displaying it in the hierarchical hypergraph module of the interactive interface; rendering the task processing schemes according to the view mode selected by the user, and displaying the rendering result in the scheme board module of the interactive interface.

[0009] According to some embodiments, the knowledge hypergraph is visualized and displayed in the hierarchical hypergraph module of the interactive interface, including: obtaining a first display graph according to the entity element layer of the knowledge hypergraph; constructing a second display graph according to the matter concept layer of the knowledge hypergraph; constructing a decision tree according to the control execution layer of the knowledge hypergraph; displaying the first display graph, the second display graph and the decision tree in the hierarchical hypergraph module of the interactive interface.

[0010] According to some embodiments, the first display graph, the second display graph and the decision tree are displayed in the hierarchical hypergraph module of the interactive interface, including: based on the hierarchical hypergraph module, highlighting the nodes of the decision tree that meet the preset conditions and labeling the reasons that meet the preset conditions; and / or based on the hierarchical hypergraph module, labeling the corresponding logical rule data and the state values of the corresponding events and entity nodes of the nodes of the decision tree for the user to click and view.

[0011] According to some embodiments, the first display graph is obtained according to the entity element layer of the knowledge hypergraph, including: constructing entity nodes according to the attribute values and state values of entities in the entity element layer of the knowledge hypergraph, and constructing edges according to the relationships between entities in the entity element layer of the knowledge hypergraph, to obtain the first display graph.

[0012] According to some embodiments, the interactive interface further includes an interactive dialogue model; in response to a scheme update instruction, updating the task processing scheme, including: in the case that the user inputs task processing demand modification information and / or task processing scheme modification information based on the interactive dialogue model, triggering the scheme update instruction; in response to the scheme update instruction, updating the task processing scheme according to the task processing demand modification information and / or the task processing scheme modification information.

[0013] According to some embodiments, in response to the scheme update instruction, the task processing scheme is updated, including: in the case that the user modifies the nodes and / or node thresholds of the decision tree based on the hierarchical hypergraph module, triggering the scheme update instruction; in response to the scheme update instruction, updating the knowledge hypergraph according to the modified nodes and / or node thresholds of the decision tree, and further updating the task processing scheme.

[0014] According to some embodiments, in response to the scheme update instruction, the task processing scheme is updated, including: in a case where the user modifies the association relationship of the entity node and / or the content of the entity node in the first display graph based on the hierarchical hypergraph module, triggering the scheme update instruction; in response to the scheme update instruction, updating the knowledge hypergraph according to the modified association relationship of the entity node and / or the content of the entity node, and further updating the task processing scheme.

[0015] According to an aspect of the present application, a task processing apparatus comprises: a knowledge hypergraph unit configured to generate a knowledge hypergraph using a preset large model according to original data and task processing prompt words; a parameter generation unit configured to generate a rule-based context parameter according to a task processing requirement input by a user; a scheme display unit configured to generate a plurality of task processing schemes using the preset large model according to the knowledge hypergraph and the context parameter, and display the task processing schemes and the knowledge hypergraph through an interactive interface; and a dynamic updating unit configured to update the task processing scheme in response to a scheme update instruction, wherein the scheme update instruction is triggered in a case where the original data is updated, the task processing requirement is modified, and / or the task processing scheme is modified, and the task processing requirement modification and the task processing scheme modification are implemented by the user based on the interactive interface.

[0016] According to an aspect of the present application, an electronic device is provided, which comprises: one or more processors; a storage device configured to store one or more programs; and when the one or more programs are executed by the one or more processors, the one or more processors implement the method as described above.

[0017] According to an aspect of the present application, a computer readable medium is provided, which stores a computer program or instructions, and the program or instructions are executed by a processor to implement the method as described above.

[0018] Through the above embodiments provided by the present application, in the hypergraph construction phase, the original data is parsed, the original data is integrated into a hypergraph structure, and the structure is used to assist the user in scheme planning, in the dynamic improvement phase, the rule-based context parameter is generated through the input of the user, the generation of the task processing scheme is started, at the same time, the scheme update instruction is generated through data update or user modification, and the task processing scheme is updated according to the scheme update instruction. The present application takes into account the automation decision efficiency and the controllability of manual intervention, can quickly generate a candidate scheme through a large model to improve the reasoning efficiency, and can also participate in the update of the scheme with the help of user interaction, solve the defects of flexibility and accuracy in pure automatic generation of the scheme, and improve the accuracy and interpretability. BRIEF DESCRIPTION OF DRAWINGS

[0019] It should be understood that the foregoing general description and the following detailed description are only exemplary and do not limit the present application.

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed in the embodiment description. Obviously, the drawings in the following description only constitute some embodiments of the present application, and other drawings can also be obtained by those skilled in the art based on these drawings without departing from the scope of the present application.

[0021] Figure 1 The flow chart of the task processing method provided by the embodiment of the present application is shown in the figure.

[0022] Figure 2 The schematic structural diagram of the three-layer hypergraph provided by the embodiment of the present application is shown in the figure.

[0023] Figure 3 The flow chart of generating multiple task processing schemes by using a preset large model according to a knowledge hypergraph and a context parameter, and displaying the task processing schemes and the knowledge hypergraph through an interactive interface is shown in the figure.

[0024] Figure 4 The flow chart of visualizing the knowledge hypergraph and displaying the layered hypergraph module in the interactive interface is shown in the figure.

[0025] Figure 5 The block diagram of the task processing device provided by the embodiment of the present application is shown in the figure.

[0026] Figure 6 The structural schematic diagram of the electronic device provided by the embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0027] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of the present application.

[0028] In addition, the described features, structures or characteristics can be combined in any suitable manner in one or more embodiments. In the following description, many specific details are provided to give a sufficient understanding of the embodiments of the present application. However, those skilled in the art will realize that the technical solutions of the present application can be practiced without one or more of the specific details, or other methods, components, devices, steps, etc. can be used. In other cases, well-known methods, devices, implementations or operations are not shown or described in detail to avoid obscuring the aspects of the present application.

[0029] The block diagrams illustrated in the drawings are merely functional entities and do not necessarily have to correspond to physically independent entities. That is, the functional entities can be implemented in the form of software, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0030] The flowcharts illustrated in the drawings are merely exemplary illustrations and do not necessarily include all contents and operations / steps, nor are they necessarily executed in the order described. For example, some operations / steps can be further decomposed, and some operations / steps can be combined or partially combined, so the actual execution order can be changed according to actual conditions.

[0031] It should be understood that although the terms first, second, third, etc. can be used herein to describe various components, these components should not be limited by these terms. These terms are used to distinguish one component from another component. Therefore, the first component discussed below can be called the second component without departing from the teachings of the present application concept. As used herein, the term "and / or" includes any one and all combinations of the associated listed items.

[0032] The specific implementation can refer to the following embodiments.

[0033] Figure 1 A flowchart of a task processing method provided by the embodiments of the present application is shown in FIG. 1. As shown in FIG. 1, the method includes steps S110-S140. Figure 1

[0034] In step S110, a knowledge hypergraph is generated by using a preset large model according to original data and a task processing prompt word.

[0035] The original data refers to multi-source data used to construct the knowledge hypergraph. The sources of the original data include but are not limited to external APIs, external sensors, databases, etc. The present application does not limit the type of original data.

[0036] In actual operation, multi-source data is obtained as original data, and then the construction of the knowledge hypergraph is performed.

[0037] Specifically, the original data is combined with the pre-set task processing prompt word to call the preset large model to perform semantic analysis and structured processing (for example, converting a natural language description of a strategy into a rule logic) on the original data, and then data fusion and hierarchical modeling are performed to obtain the knowledge hypergraph.

[0038] Further, in some embodiments, for some ambiguous information, the large model prompts the user to disambiguate.

[0039] ​The task processing prompt word is a prompt word constructed according to a task to be processed, and contains entities, logical rules, and the like.

[0040] The preset large model can be a public large model, or can be constructed according to a sample data set of a task to be processed, and the large model is trained by using the sample data set, and the present application does not limit this.

[0041] It needs to be explained that the knowledge supergraph refers to a knowledge graph used to describe event logic, state changes, and causal relationships, and focuses on event evolution rules and rule reasoning.

[0042] In step S120, a context parameter of a rule is generated according to a task processing requirement input by a user.

[0043] In the dynamic improvement stage, first, a user inputs a task processing requirement, generates a context parameter of a rule, and starts generation of a task scheme.

[0044] The user task processing requirement can be input through an interactive interface.

[0045] In step S130, a plurality of task processing schemes are generated by using a preset large model according to a knowledge supergraph and a context parameter, and the task processing schemes and the knowledge supergraph are displayed through an interactive interface.

[0046] After the knowledge supergraph and the context parameter are obtained, generation of a task scheme is started.

[0047] In the calculation process, a plurality of task processing schemes are generated according to the knowledge supergraph and the context parameter in combination with a supergraph model in a database, and the task processing schemes and the knowledge supergraph are pushed through an interactive interface. A user can view the task processing schemes and the knowledge supergraph through the interactive interface.

[0048] Further, in some embodiments, the knowledge supergraph is displayed in the form of a knowledge tree.

[0049] According to the example embodiments, the present application will push a plurality of schemes (part or all of the generated task processing schemes) that meet default constraint conditions by default, and a user can select an existing scheme or add a new scheme.

[0050] Further, the interactive interface also supports operations of a user for scheme query, scheme selection, and scheme optimization.

[0051] In step S140, a task processing scheme is updated in response to a scheme update instruction, wherein the scheme update instruction is triggered in a case of original data update, task processing requirement modification, and / or task processing scheme modification, and the task processing requirement modification and the task processing scheme modification are implemented by a user based on an interactive interface.

[0052] On the basis of the content displayed in the interactive interface, the user can view the specific information of the task processing scheme and the specific content of the knowledge supergraph, such as rule constraints, events, and entity states.

[0053] The user can modify the task processing demand or the task processing scheme according to the content displayed in the interactive interface, thereby triggering a scheme update instruction. The scheme update instruction can also be triggered by original data update.

[0054] Further, in some embodiments, the data changes of the original data are continuously monitored, the update is pushed, and then incremental recalculation is performed to update the supergraph state in the database and synchronize to the interactive interface. The interface automatically refreshes the view, and the intelligent assistant prompts the changes through a pop-up window / highlight.

[0055] After triggering the scheme update instruction, the task processing scheme is updated according to the content modified by the user.

[0056] According to the technical scheme of the present application, in the supergraph construction stage, the original data is parsed, and the original data is integrated into a supergraph structure. Based on the structure, the user is assisted in scheme planning. In the dynamic improvement stage, the context parameters of the rules are generated through the input of the user, the generation of the task processing scheme is started, and at the same time, the scheme update instruction is generated through data update or user modification. The task processing scheme is updated according to the scheme update instruction. The present application takes into account the efficiency of automatic decision and the controllability of manual intervention. It can quickly generate candidate schemes through a large model to improve the reasoning efficiency, and at the same time, it can participate in the update of the scheme with the help of user interaction, solve the defects of flexibility and accuracy in pure automatic generation of the scheme, and improve the accuracy and explainability.

[0057] According to some embodiments, the knowledge supergraph includes: an entity element layer for storing attribute values and state values of entities and relationships between entities; a matter concept layer for storing logical rule data, wherein the logical rule data is used to describe the trigger conditions and / or logic of events; and a control execution layer for generating an execution scheme according to the logical rule data and the state values of the entities.

[0058] Specifically, in this embodiment, the knowledge supergraph is a three-layer supergraph structure obtained by integrating multi-source heterogeneous data. Specifically, it includes a matter concept layer (matter rule network), a control execution layer (control node network), and an entity element layer (entity knowledge network); and at the same time, it assists humans in scheme planning and dynamic decision-making based on the structure.

[0059] Among them, the affair concept layer constructs an inferable event network based on the multi-dimensional association (causal, chronological, conditional dependence, etc.) between events and possible event state changes, and stores rules for describing event trigger conditions and logic; the control execution layer is the core of the entire task processing framework, responsible for generating executable decision paths, supporting dynamic condition judgment and branch selection, generating specific execution plans, and dynamically adjusting the plan combination according to the changes of events. The entity element layer is the specific embodiment of the affair graph, which stores the attribute values and state values of all entity nodes and the relationship between entities, serving as the underlying data support.

[0060] In actual operation, between layers, the entity element layer provides the relationship and state of the time event to the affair concept layer; the affair concept layer provides the relationship and state of the time event to the control execution layer and transmits rules to guide the control execution layer to handle the changes of event state; the control execution layer generates specific plans according to the rule trigger conditions combined with the entity state of the entity element layer, and realizes the adjustment and optimization of the plans according to the changes of event state, rules or entity state values; the control execution layer calls the data of the entity element layer and writes the entity state changes into the entity element layer; the affair concept layer is an abstraction of the entity element layer, providing semantic expression of domain knowledge. Figure 2 The schematic structure of the three-layer hypergraph is given in the specification.

[0061] The application introduces an auxiliary computing background as the core hub to coordinate the human-computer interaction interface, large model service and external service.

[0062] The application adopts the three-layer hypergraph architecture of the affair concept layer, the control decision layer and the entity element layer, abstracts the affair concept and rules in a specific scenario, associates the event state with the entity attribute and state of the entity element layer through the control decision layer, realizes independent management within the layer and dynamic changes across layers, solves the coupling problem of traditional single-layer graph model in complex constraint processing, and supports multi-dimensional decision-making in complex scenarios.

[0063] According to some embodiments, the task processing prompt words include entity description templates of the entity element layer, event description templates of the affair concept layer, behavior tree framework templates of the control execution layer, and output sample examples.

[0064] In this embodiment, the task processing prompt words include templates of the three-layer hypergraph that can be reused across domains, and the templates include definitions and requirement explanations of each layer.

[0065] Specifically, the task processing prompt words include: entity description templates of the entity element layer, event description templates of the affair concept layer, behavior tree framework templates of the control execution layer, and output sample examples.

[0066] In one specific embodiment, the task processing prompt words include:

[0067] {"event_templates":{"atomic_events":[{"event_id":"EVT_APT","name":"Reservation {venue type}","type":"atomic event","params":["venue type","{city}"],...”"composite_events":[{"event_id":"EVT_CULTURE","name":"{city}cultural experience","type":"composite event","sub_events":["EVT_APT[venue type = museum]", ... T_TRAVEL[method = walking]"],"relations":{"cause and effect":"Reservation successful → Tour can begin","sequence":"Transportation requires {buffer time} before the visit","mutual exclusion":"Other {activity types} cannot be scheduled during the same time period"}}],"domain_events":[{"event_id":"EVT_WEATHER","name":"Weather change event","impact_layers":["reasoning","control","entity"],"rule_template":"if {weather type}then adjust threshold = {dynamic threshold}"}]}

[0068] "behavior_tree_framework":{"core_nodes":{"Selector":{"logic":"Try branches according to priority until success","data_binding":"{city}.Featured Activities List","params":["Priority Strategy"]},"DynamicInject":{"type":"Dynamic Node","trigger_condition":"Event Rule.{Rule ID} Trigger","action":"Insert {city}.Featured Activities Template"}},"rule_binding":{"weather_rule":{"condition":"Entity Layer.Weather Condition ∈ {Severe Weather List}","action":"Activate Alternative Branch:indoor_activities"}}}

[0069] "entity_templates": { "place": { "attributes": ["type", "coordinates", "{city} signature tags"], "states": ["bookable", "temporary closure", "crowd alert"], "relations": ["adjacent_{transportation type}"]}, "transportation": { "attributes": ["mode", "real-time delay index"], "state_machine": { "states": ["normal", "partial disruption", "complete disruption"], "threshold_params": ["{delay threshold}"]}}, "clothing": { "attributes": ["temperature range", "{season} adaptability"]}}

[0070] According to some embodiments, referring to Figure 3 In step S130, according to the knowledge hypergraph and the context parameters, a plurality of task processing schemes are generated by using a preset large model, and the task processing schemes and the knowledge hypergraph are displayed through an interactive interface. Specifically, steps S310-S330 can be implemented.

[0071] In step 310, according to the context parameters, a plurality of task processing schemes are generated by using a preset large model, based on the knowledge hypergraph, to adjust the execution scheme of the control execution layer.

[0072] After obtaining the context parameters input by the user, the context completion can be performed according to the context parameters, so as to generate a plurality of recommended schemes in combination with the knowledge hypergraph in the database.

[0073] The control execution layer can generate a specific scheme according to a rule trigger condition in combination with an entity state of the entity element layer, and realize adjustment and optimization of the scheme according to changes of an event state, a rule or an entity state value. That is, in the initial hypergraph generation stage, the control execution layer generates an execution scheme according to a rule trigger condition in combination with an entity state of the entity element layer. In the dynamic improvement stage, the user performs operations such as context completion and event activation, and the control execution layer realizes adjustment and optimization of the scheme according to changes of an event state, a rule or an entity state value, thereby generating a plurality of task processing schemes.

[0074] In step 320, the knowledge hypergraph is visualized and displayed in the layered hypergraph module of the interactive interface.

[0075] The layered hypergraph module is constructed in the interactive interface, and the knowledge hypergraph is displayed.

[0076] According to the example embodiment, the affair concept layer and the entity element layer are displayed in the form of a graph, and the control decision layer is displayed in the form of a decision tree.

[0077] In step 330, the task processing scheme is rendered according to the view mode selected by the user, and the rendering result is displayed on the scheme board module of the interactive interface.

[0078] The generated multiple task processing schemes are visualized and provided for the user to select, and the scheme board module is constructed to display the same.

[0079] The user can view the details of the scheme and the generated state dependency and rule constraints through the scheme board module. In actual operation, the user can select a view mode (such as a timeline view or a map view) from a view template library, and the scheme board module dynamically renders the details of the task processing scheme based on the view mode and the knowledge hypergraph to obtain a rendering result for the user to view and select.

[0080] The view mode can be set according to requirements, and the present application does not limit this.

[0081] According to an example embodiment, a scheme selector is constructed to support the user to select a task processing scheme.

[0082] Further, in some embodiments, the user can also perform conflict resolution operations (adjust rules, event states, etc.) or achieve dynamic optimization of the scheme by adjusting rule weights based on the scheme board module.

[0083] Based on the above embodiments, after the scheme optimization is completed, the user can still select a view mode from the view template library to re-render the task processing scheme.

[0084] According to some embodiments, with reference to Figure 4 In step 320, the knowledge hypergraph is visualized and displayed on the layered hypergraph module of the interactive interface, which can be implemented through steps S410-S440.

[0085] In step S410, a first display graph is obtained according to the entity element layer of the knowledge hypergraph.

[0086] In step S420, a second display graph is constructed according to the affair concept layer of the knowledge hypergraph.

[0087] In step S430, a decision tree is constructed according to the control execution layer of the knowledge hypergraph.

[0088] In step S440, the first display graph, the second display graph, and the decision tree are displayed on the layered hypergraph module of the interactive interface.

[0089] The entity element layer is rendered in the form of a graph as the first display graph, the affair concept layer is rendered in the form of a graph as the second display graph, and the control decision layer is displayed in the form of a tree decision tree.

[0090] According to some embodiments, in step S440, the first display graph, the second display graph and the decision tree are displayed in the hierarchical hypergraph module of the interactive interface, which can be implemented through step S441 and / or step S442.

[0091] In step S441, based on the hierarchical hypergraph module, the nodes of the decision tree that meet the preset condition are highlighted and displayed, and the reason for meeting the preset condition is marked.

[0092] The nodes that meet the preset condition are highlighted, so that the user can view the specific reason.

[0093] According to an example embodiment, for the conflict nodes, the specific conflict reason is marked in the form of highlighting, and the user can view the specific conflict reason.

[0094] In step S442, based on the hierarchical hypergraph module, the nodes of the decision tree are marked with the corresponding logical rule data and the state values of the corresponding event and entity nodes for the user to click and view.

[0095] The nodes of the decision tree are marked with the logical rule data and the state values of the corresponding event and entity nodes, so that the user can view the specific rule constraints, events and entity states by clicking the nodes in the tree.

[0096] According to some embodiments, in step S410, the first display graph is obtained according to the entity element layer of the knowledge hypergraph, specifically including: constructing entity nodes according to the attribute values and state values of entities in the entity element layer of the knowledge hypergraph, and constructing edges according to the relationships between entities in the entity element layer of the knowledge hypergraph, to obtain the first display graph.

[0097] According to some embodiments, the interactive interface further includes an interactive dialogue model.

[0098] Based on the above embodiments, in step S140, in response to the scheme update instruction, the task processing scheme is updated, which can be implemented through step S141 to step S142.

[0099] In step S141, in the case where the user inputs task processing requirement modification information and / or task processing scheme modification information based on the interactive dialogue model, the scheme update instruction is triggered.

[0100] That is, the human-computer interaction interface includes a hierarchical view module, a scheme dashboard module and an interactive dialogue model. The user can perform inquiry information completion or conflict adjustment through the interactive dialogue model.

[0101] This embodiment is described taking an intelligent chat assistant as an example of the interactive dialogue model, but it does not represent a limitation on the present application.

[0102] According to an example embodiment, the user inputs task processing demand modification information and / or task processing scheme modification information in the chat window of the intelligent chat assistant to make inquiring information supplement or conflict adjustment, and triggers a scheme update instruction.

[0103] Further, in some embodiments, the user modifies conditions or associated entities in the decision tree corresponding to the current scheme, and in the case of conflict, the intelligent chat assistant prompts the user in the chat window. The user inputs task processing demand modification information and / or task processing scheme modification information in the chat window of the intelligent chat assistant to make inquiring information supplement or conflict adjustment, and triggers a scheme update instruction.

[0104] According to an example embodiment, the matching of the user context data mapping and the rules is detected in real time to determine whether the current scheme has a conflict. In the case of conflict, a prompt is popped up in the chat window to give a warning.

[0105] In step S142, in response to the scheme update instruction, the task processing scheme is updated according to the task processing demand modification information and / or the task processing scheme modification information.

[0106] According to the scheme update instruction, the task processing scheme is updated, and natural language driven scheme adjustment is realized.

[0107] By integrating multi-source data (text, geographic location information, real-time data, etc.), the static limitations of traditional rule engines are broken through, and data-driven rule dynamic changes are realized, making the planning scheme more time-effective and scene-adaptive. The advantage is that it takes into account the efficiency of automatic decision-making and the controllability of manual intervention. It can quickly generate candidate schemes through large models, and ensure rule logic transparency and conflict resolution traceability through visual interaction, significantly improving the consistency of schemes in complex decision-making scenarios.

[0108] According to some embodiments, in step S140, in response to the scheme update instruction, the task processing scheme is updated, which can be realized by steps S143-S144.

[0109] In step S143, in the case where the user modifies the nodes and / or node thresholds of the decision tree based on the hierarchical hypergraph module, the scheme update instruction is triggered.

[0110] Based on the decision tree displayed by the hierarchical hypergraph module, the user can adjust the details of the decision tree according to the prompt, and trigger the scheme update instruction.

[0111] The adjustment of the details of the decision tree includes modifying the nodes, node thresholds, etc. of the decision tree.

[0112] The prompt is a prompt given by the interaction model in the case of conflict between the context data mapping and the rules.

[0113] In step S144, in response to the scheme updating instruction, the knowledge hypergraph is updated according to the modified nodes and / or node thresholds of the decision tree, and the task processing scheme is further updated.

[0114] The knowledge hypergraph is updated according to the scheme updating instruction, and the task processing scheme is further updated, so as to realize the customer adjustment driven scheme adjustment.

[0115] According to some embodiments, in step S140, in response to the scheme updating instruction, the task processing scheme is updated, which can be realized by steps S145-S146.

[0116] In step S145, in the case that the user modifies the association relationship of the entity node and / or the content of the entity node in the first display graph based on the hierarchical hypergraph module, the scheme updating instruction is triggered.

[0117] Based on the first display graph displayed by the hierarchical hypergraph module, the user can make detailed adjustments to the first display graph to trigger the scheme updating instruction.

[0118] The adjustment of the first display graph includes modifying the input parameters, dragging the association relationship of the entity node, adjusting the rule weight, etc.

[0119] In step S146, in response to the scheme updating instruction, the knowledge hypergraph is updated according to the modified association relationship of the entity node and / or the content of the entity node, and the task processing scheme is further updated.

[0120] The knowledge hypergraph is updated according to the scheme updating instruction, and the task processing scheme is further updated, so as to realize the user adjustment driven scheme adjustment.

[0121] The present application integrates the user into the man-machine collaborative decision of the decision closed loop, and completes the scheme making through conflict visualization and progressive guidance. An interactive interface is provided, and the user can intuitively adjust the combination relationship of the scheme elements and the weight parameter of the scheme rule; through the interactive means, the defects of flexibility and accuracy in the pure automatic generation of the scheme are solved. At the same time, a prefabricated visualization template is provided to generate a view more in line with the user's habits, and the practicability and user satisfaction of the solution scheme are improved.

[0122] Based on the above embodiment, the human-computer interaction interface includes a hierarchical view module, a scheme dashboard module, and an interactive dialogue model. The hierarchical hypergraph module can display summary information or detailed information of the three-layer hypergraph structure according to user needs, support semantic focusing, for example, highlighting the current construction of the semantic related nodes, highlighting the conflict nodes, etc. The scheme dashboard module provides a scheme selector and a visualization plug-in (figure, table, tree, map, etc.) for scheme display. Users can customize the scheme view template according to their needs to realize the visualization display of the scheme or scheme details. The interactive dialogue model supports inquiry-based information completion or conflict adjustment to realize natural language-driven scheme adjustment. On this basis, on the one hand, the accuracy of the hypergraph construction is guaranteed through visual interactive verification and constraint injection, and on the other hand, the internal relationship of the complex structure is better understood through hierarchical display, semantic focusing, and other visualization means to assist users in planning the scheme.

[0123] In one specific embodiment, after entering the dynamic decision-making phase, the view of the interactive interface is divided into two columns, including the scheme dashboard module and the chat assistant. By default, a number of schemes that meet the default constraint conditions are pushed, and the user can select an existing scheme or add a new scheme, modify conditions in the decision tree corresponding to the current scheme, or associate entities. For cases that do not meet the conditions, a prompt is given in the chat window.

[0124] Based on the above embodiment, after all conflicts in the current scheme are resolved, the user can select a suitable view through "view or export" to fill the corresponding data into the preset template to generate the expected view for the user. If the template cannot meet the needs, the user can add a new template according to the needs.

[0125] The present application introduces an auxiliary computing background as the core hub to coordinate the human-computer interaction interface, the large model service, and the external service. First, it analyzes the original data and generates a three-layer hypergraph structure based on human semantics. Second, it detects user context data mapping and rule conflicts in real time and gives early warnings through the chat window. At the same time, it monitors entity state values through external services to trigger the dynamic re-planning of the rule engine. Third, it pushes the newly generated scheme to the scheme dashboard of the human-computer interaction interface to support user operations such as scheme query, scheme selection, and scheme optimization.

[0126] In order to provide a more detailed description of the task processing method provided by the present application, the following specific embodiment is given based on the multi-objective optimization problem commonly faced by users in the field of tourism planning.

[0127] It can be understood that the traditional method relies on static rules or single data source, and it is difficult to handle dynamic data and complex semantic relationship. Taking "economic parent-child three-day tour" as an example, it is necessary to integrate heterogeneous data such as travel time, scenic spot opening time, transportation cost, hotel comfort, etc. However, the existing technology has three major pain points of data fragmentation, rigid rules and cognitive overload. Scenic spot information, real-time traffic and user preferences are scattered in different platforms, and preset constraints cannot adapt to dynamic changes such as weather changes or traffic delays, and complex relationship network makes it difficult for users to quickly locate the conflict source.

[0128] Based on the task processing method provided in the present application, a hierarchical hypergraph architecture is proposed, which includes five core modules of heterogeneous data source access, hypergraph construction engine, storage module, adaptive feedback module and human-computer interaction interface. Through dynamic data fusion and rule adaptive optimization, accurate modeling is realized - element layer integrates entity nodes such as scenic spots, transportation and hotels, rule layer defines logical constraints such as budget and time, and scheme layer aggregates to generate executable itinerary. The whole interaction interface supports conflict tracing, dynamic adjustment and multi-objective optimization.

[0129] In the initial generation stage of hierarchical hypergraph, the heterogeneous data module sets up "tourist strategy analysis Agent", sets up target website (Little Red Book, Mavens, etc.), crawling strategy, etc., to obtain original data.

[0130] In the large model and initial hypergraph construction stage, the prompt words of the large model include entities and logical rules, and the scheme involves topics, scenic spots, accommodation, travel and catering.

[0131] #Large model prompt words

[0132] prompt="""

[0133] Please generate a set of cross-domain reusable three-layer hypergraph templates based on the following requirements:

[0134] **Requirement Description**:

[0135] 1.**Concept layer of matter and reason**: Define general event templates, including atomic events (not divisible), composite events (multi-step combination), and domain events (cross-layer influence). It needs to meet:

[0136] -Event parameterization (such as `{place type}`, `{city}`).

[0137] -Relationship between events includes cause and effect, time sequence, and mutual exclusion.

[0138] -Rule template supports dynamic injection threshold (such as weather sensitivity, crowd threshold).

[0139] 2.**Control execution layer**: Design a behavior tree framework, which requires:

[0140] - Core nodes (Selector / Sequence / Parallel) can dynamically load city data.

[0141] - Branch logic is bound to the rules layer (e.g., `Weather Rule → Trigger Alternative Branch`).

[0142] - Contains "Dynamic Injection Nodes" for inserting city-specific activities.

[0143] 3. **Entity Element Layer**: Defines entity type templates, including attributes, states, relationship fields, and supports city instantiation.

[0144] **Output Format**: Strict JSON structure, containing the following keys:

[0145] - `event_templates` (Event Template Library)

[0146] - `behavior_tree_framework` (Behavior Tree Framework)

[0147] - `entity_templates` (Entity Type Templates)

[0148] **Constraints**:

[0149] - Specific instances (e.g., "Forbidden City", "Yamanote Line") are prohibited; only abstract placeholders are allowed.

[0150] - State machines must include `Not Started / Ready / In Progress / Completed / Failure` states.

[0151] - Each template must be labeled with extensible parameters (represented by `{}`).

[0152] """

[0153] Large model output parses similar results:

[0154] {"event_templates":{"atomic_events":[{"event_id":"EVT_APT","name":"Reservation {venue type}","type":"atomic event","params":["venue type","{city}"],...”"composite_events":[{"event_id":"EVT_CULTURE","name":"{city}cultural experience","type":"composite event","sub_events":["EVT_APT[venue type = museum]", ... T_TRAVEL[method = walking]"],"relations":{"cause and effect":"Reservation successful → Tour can begin","sequence":"Transportation requires {buffer time} before the visit","mutual exclusion":"Other {activity types} cannot be scheduled during the same time period"}}],"domain_events":[{"event_id":"EVT_WEATHER","name":"Weather change event","impact_layers":["reasoning","control","entity"],"rule_template":"if {weather type}then adjust threshold = {dynamic threshold}"}]}

[0155] "behavior_tree_framework":{"core_nodes":{"Selector":{"logic":"Try branches according to priority until success","data_binding":"{city}.Featured Activities List","params":["Priority Strategy"]},"DynamicInject":{"type":"Dynamic Node","trigger_condition":"Event Rule.{Rule ID} Trigger","action":"Insert {city}.Featured Activities Template"}},"rule_binding":{"weather_rule":{"condition":"Entity Layer.Weather Condition ∈ {Severe Weather List}","action":"Activate Alternative Branch:indoor_activities"}}}

[0156] "entity_templates": {"place": {"attributes": ["type", "coordinates", "{city} signature tags"], "states": ["bookable", "temporary closure", "crowd alert"], "relations": ["adjacent_{transport type}"]}, "transport": {"attributes": ["mode", "real-time delay index"], "state_machine": {"states": ["normal", "partial disruption", "complete disruption"], "threshold_params": ["{delay threshold}"]}}, "clothing": {"attributes": ["temperature range", "{season} adaptability"]}}

[0157] Based on the above template, the data parser of the auxiliary intelligent background calls the large model to extract information in the travel guide based on the template. In the extraction process, threshold values are set. For some ambiguous information or information that cannot meet the template, the intelligent assistant is fed back to the chat window in the human-computer interface, prompting the user to assist in processing.

[0158] After processing, the hypergraph calculation module in the auxiliary intelligent background processes the parsed structured information into three layers: 1) travel-related event and rule network, including atomic events (such as accommodation / transportation booking), composite events (such as daily itinerary), and domain events (such as weather changes), which are constructed by defining the logical relationship between events and evolution rules; 2) decision tree based on event state and rules; 3) travel-related entities (such as scenic spots, restaurants, hotels, etc.) and their relationship itinerary network. The information is rendered through the human-computer interface. Users can optimize the three layers of information or adjust the decision tree structure through the intelligent assistant, add entity nodes, or view detailed information of a certain layer. The hypergraph calculation module synchronously modifies the layered hypergraph information. Finally, the generated knowledge hypergraph is displayed in the layered view module.

[0159] In the planning stage, the application preforms some plans, users can choose existing plans as templates according to their own needs, or re-plan, in this embodiment, the "plan planning" interface is divided into two parts, one part is the plan board module, and the other part is the chat assistant. The plan board module is divided into two parts, one part is the decision tree, which corresponds to the decision tree instance generated by the control execution layer of the hierarchical hypergraph structure, and the right side is the schedule panel. Users can edit the decision tree to add or modify rules (such as adjusting the upper limit of the budget, adding or adjusting user preferences, etc.), adjust or select specific entities (such as modifying hotels, restaurants, etc.); You can also directly chat with the chat assistant to express your modification intention. The hypergraph calculation module of the auxiliary decision layer will synchronize the calculation, query whether there is a conflict in the current plan, if there is a conflict, the chat assistant and the plan panel will be synchronized to render the conflict reason and prompt the user to make a new selection.

[0160] In the actual execution process, first, based on the existing "luxury travel plan", the chat assistant prompts the user to complete the basic information of the trip to meet the "travel necessary rules cannot be empty" rule built in the decision tree. After the user completes the information through the chat assistant, replace the entities such as restaurants and hotels in the decision tree, and push the budget exceeding the limit to the chat assistant in the background. The user needs to adjust the involved entity unit until it meets the budget or directly modifies the budget. Similarly, there may be time conflicts, preference conflicts, and other scenarios, such as the user-selected scenic spot is closed during the user's trip, or the user-selected flight has a high chance of delay, causing conflicts with subsequent events. You can give the user a prompt according to the priority to help the user understand the internal reason for the conflict and modify the relevant content until the plan is completed.

[0161] In the actual tourism planning process, the plan generated for the user needs to consider the understandability of the view. Through the time sequence triggered by the event state associated by the decision tree, the schedule planning table is used to represent the schedule, or the trajectory graph can be drawn with the help of the geographical location parameters of different types of entities, or the aggregated subclass view can be generated by using the attribute values or state values shared by different types of entities, such as cost analysis, travel physical consumption analysis, etc. At the same time, the view can be dynamically updated after the user modifies the entity attributes or adjusts the plan.

[0162] For example, the pre-set schedule trajectory template contains a map component, and the hypergraph calculation module aggregates the time and location information of all element entities contained in the current plan to complete the template data filling. Finally, the entity is marked on the coordinate position on the final view, and the trajectory is drawn according to the time sequence to form a date directed graph.

[0163] A pre-set budget tree structure, the initial view shows that the budget information is contained above the plan layer card, and clicking the Modal box displays a three-layer tree structure, the first layer is the total budget, the second layer is the classification information, and the third layer is the element information. For example:

[0164] Economic three-day tour (total cost 4800)

[0165] ├─Sights: Terracotta Warriors (120), Huaqing Pool (80)

[0166] ├─Transportation: Subway Line 2 (10 yuan x 3 days)

[0167] └─Hotel: Family-friendly hotel (300 yuan x 3 nights)

[0168] Further, the built-in monitoring module obtains information through regular active pulling or message subscription and synchronously modifies the entity attribute values of the element layer. Changes in entity attribute values may violate the constraints formed in the scheme, making the scheme unexecutable. For example, the weather API pushes "XX tomorrow's rainstorm", automatically marking the element layer entity node outdoor scenic spot as "temporary closure", and the flight as "90% delay". The interactive interface pop-up window prompts "the current travel scheme is invalid", the intelligent assistant prompts the invalid reason, and the element layer entity becomes gray. The user needs to reselect the entities associated with the scheme, adjust the rule strategy, and complete the re-formulation of the scheme.

[0169] In addition to the restrictions of hard rules, users can also adjust the scheme by adjusting the weights of soft rules. For example, set the theme node "family-friendly" in the scheme layer, and add a new rule function to filter entities containing the "family-friendly" feature. If the current entity has no attributes, the user can directly operate the element layer elements, such as adding a zoo or science museum. The newly generated scheme may have multiple options, and the user can adjust the priority of the alternative scheme by adjusting the weight of the "family-friendly rule".

[0170] For example, the user can set the rule weights of the scheme on the operation panel: cost rule (0.6), time rule (0.3), and comfort rule (0.1). At this time, three schemes are output:

[0171] 1) Economic type (cost 4500 yuan, time 6h / day, comfort 3.8);

[0172] 2) Balanced type (cost 5000 yuan, time 4h / day, comfort 4.2);

[0173] 3) Comfort type (cost 5800 yuan, time 3h / day, comfort 4.7).

[0174] The user can change "car service" in the "balanced type" scheme to "subway + shared bicycle" in the decision tree or scheme view, and the hypergraph calculation module calculates the new parameters in real time (cost ↓4800 yuan, time ↑5.5h / day).

[0175] The current use case is used to illustrate the application of the present application in the application example of the tourism planning scene. The task processing method provided by the present application can be applied to the fields of medical treatment, logistics and the like. For example, in the medical treatment scene, by integrating the entity element layer to integrate hospital, drug and medical staff data, the control decision layer defines the disease diagnosis and treatment process, the event relationship and the corresponding constraint of the treatment operation and treatment path in the event concept layer, and finally becomes an auxiliary medical treatment assistant. The core technologies such as hierarchical hypergraph structure and man-machine collaborative feedback mechanism promote cross-industry knowledge modeling and decision optimization, realize the deep integration of artificial intelligence and field experience, and thus provide flexible, transparent and interpretable intelligent support for complex scenes.

[0176] The device embodiment of the present application is described below, which can be used to execute the method embodiment of the present application. For details not disclosed in the device embodiment of the present application, reference can be made to the method embodiment of the present application.

[0177] Figure 5 A block diagram of a task processing device according to an example embodiment is shown.

[0178] Figure 5 The device shown can execute the task processing method according to the embodiments of the present application as described above.

[0179] As shown in Figure 5 The task processing device can include:

[0180] The knowledge hypergraph unit 510 is configured to generate a knowledge hypergraph by using a preset large model according to original data and a task processing prompt word.

[0181] The parameter generation unit 520 is configured to generate a rule context parameter according to a task processing requirement input by a user.

[0182] The scheme display unit 530 is configured to generate a plurality of task processing schemes by using a preset large model according to the knowledge hypergraph and the context parameter, and display the task processing schemes and the knowledge hypergraph through an interactive interface.

[0183] The dynamic updating unit 540 is configured to update the task processing scheme in response to a scheme updating instruction, wherein the scheme updating instruction is triggered in the case of original data updating, task processing requirement modification and / or task processing scheme modification, and the task processing requirement modification and the task processing scheme modification are realized by the user based on the interactive interface.

[0184] The device performs similar functions to the method provided above, and other functions can be referred to the foregoing description, which will not be described here again.

[0185] The embodiment of the present application discloses an electronic device, comprising: a processor; a memory storing a computer program, when the computer program is executed by the processor, the processor executes the above instruction generation method.

[0186] For example, referring to the electronic device 600 shown in FIG. 6A, the electronic device 600 includes a processor 601 and a memory 603. The processor 601 and the memory 603 are connected, for example, via a bus 602. Optionally, the electronic device 600 can further include a transceiver 604. It should be noted that the transceiver 604 is not limited to one in actual applications, and the structure of the electronic device 600 does not constitute a limitation on the embodiments of the present application. Figure 6 , Figure 6 The processor 601 can be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. The processor 601 can implement or execute various exemplary logical blocks, modules, and circuits described in the disclosure. The processor 601 can also be a combination of computing functions, such as one or more microprocessor combinations, combinations of DSP and microprocessor, etc.

[0187] The bus 602 can include a path for transmitting information between the above-mentioned components. The bus 602 can be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus 602 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but this does not mean that there is only one bus or only one type of bus.

[0188] The bus 602 can include a path for transmitting information between the above-mentioned components. The bus 602 can be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus 602 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but this does not mean that there is only one bus or only one type of bus. Figure 6

[0189] ​The memory 603 can be a ROM (Read Only Memory) or other type of static storage device that can store static information and instructions; a RAM (Random Access Memory) or other type of dynamic storage device that can store information and instructions; an EEPROM (Electrically Erasable Programmable Read-Only Memory), a CD-ROM (Compact Disc Read-Only Memory) or other optical disk storage; a magnetic disk storage or other magnetic storage devices; or any other medium that can be used to carry or store desired program codes in the form of instructions or data structures and that can be accessed by a computer, but is not limited to this.

[0190] The memory 603 is configured to store application program codes for implementing the solutions of the present application, and the processor 601 is configured to control the execution of the application program codes. The processor 601 is configured to execute the application program codes stored in the memory 603 to implement the content shown in the foregoing method embodiments.

[0191] Figure 6 The electronic device shown is only an example, and should not impose any limitation on the functions and use range of the embodiments of the present application.

[0192] The embodiments of the present application disclose a computer readable storage medium, which stores a computer program. When the computer program is executed by a processor, the processor executes an instruction generation method.

[0193] It should be understood that, although each step in the flowchart of the accompanying drawings is shown in sequence according to the direction of the arrow, these steps are not necessarily executed in sequence according to the direction of the arrow. Unless otherwise specified herein, the execution of these steps is not strictly limited in sequence, and they can be executed in other sequences. Moreover, at least part of the steps in the flowchart of the accompanying drawings can include multiple sub-steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or sub-steps or stages of other steps.

[0194] The above is only some embodiments of the present application, and it should be pointed out that, for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, and these improvements and refinements should also be considered as the protection scope of the present application.

Claims

1. A task processing method, characterized in that, include: Based on the raw data and task processing prompts, a knowledge hypergraph is generated using a pre-set large model. Generate context parameters for the rules based on the user's input of task processing requirements; Based on the knowledge hypergraph and the context parameters, multiple task processing schemes are generated using the preset large model, and the task processing schemes and the knowledge hypergraph are displayed through an interactive interface; In response to a scheme update command, the task processing scheme is updated, wherein the scheme update command is triggered when the original data is updated, the task processing requirements are modified, and / or the task processing scheme is modified, and the modification of the task processing requirements and the modification of the task processing scheme are implemented by the user based on the interactive interface.

2. The method according to claim 1, characterized in that, The knowledge hypergraph includes: The entity element layer is used to store the attribute values ​​and state values ​​of entities, as well as the relationships between the entities; The event concept layer is used to store logical rule data, wherein the logical rule data is used to describe the triggering conditions and / or logic of events; The control execution layer is used to generate an execution plan based on the logical rule data and the state value of the entity.

3. The method according to claim 2, characterized in that, The task processing prompts include entity description templates for the entity element layer, event description templates for the event concept layer, behavior tree framework templates for the control execution layer, and output sample examples.

4. The method according to claim 2, characterized in that, Based on the knowledge hypergraph and the context parameters, multiple task processing schemes are generated using the preset large model, and the task processing schemes and the knowledge hypergraph are displayed through an interactive interface, including: Based on the context parameters, using the preset large model and the knowledge hypergraph, the execution scheme of the control execution layer is adjusted to generate multiple task processing schemes; The knowledge hypergraph is visualized and displayed in the hierarchical hypergraph module of the interactive interface; Based on the view mode selected by the user, the task processing plan is rendered, and the rendering result is displayed in the plan dashboard module of the interactive interface.

5. The method according to claim 4, characterized in that, The knowledge hypergraph is visualized and displayed in the hierarchical hypergraph module of the interactive interface, including: Based on the entity element layer of the knowledge hypergraph, a first display diagram is obtained; Based on the concept layer of the knowledge hypergraph, a second display graph is constructed; Construct a decision tree based on the control execution layer of the knowledge hypergraph; The first display diagram, the second display diagram, and the decision tree are displayed in the hierarchical supergraph module of the interactive interface.

6. The method according to claim 5, characterized in that, Displaying the first display image, the second display image, and the decision tree in the hierarchical hypergraph module of the interactive interface includes: Based on the hierarchical hypergraph module, nodes of the decision tree that meet preset conditions are highlighted and the reasons for meeting the preset conditions are annotated; and / or Based on the hierarchical hypergraph module, the nodes of the decision tree are labeled with corresponding logical rule data and their corresponding event and entity node status values, so that users can click to view them.

7. The method according to claim 5, characterized in that, Based on the entity element layer of the knowledge hypergraph, a first display graph is obtained, including: Entity nodes are constructed based on the attribute values ​​and state values ​​of the entities in the entity element layer of the knowledge hypergraph, and edges are constructed based on the relationships between the entities in the entity element layer of the knowledge hypergraph to obtain a first display graph.

8. The method according to claim 5, characterized in that, The interactive interface also includes an interactive dialogue model. In response to the scheme update instruction, the task processing scheme is updated, including: When a user inputs information to modify task processing requirements and / or task processing solutions based on the interactive dialogue model, the solution update instruction is triggered. In response to the scheme update instruction, the task processing scheme is updated according to the task processing requirement modification information and / or task processing scheme modification information.

9. The method according to claim 5 or 6, characterized in that, In response to the scheme update instruction, the task processing scheme is updated, including: When a user modifies the nodes and / or node thresholds of the decision tree based on the hierarchical hypergraph module, the scheme update instruction is triggered. In response to the scheme update instruction, the knowledge hypergraph is updated according to the modified nodes and / or node thresholds of the decision tree, thereby updating the task processing scheme.

10. The method according to claim 7, characterized in that, In response to the scheme update instruction, the task processing scheme is updated, including: When a user modifies the relationships and / or content of entity nodes in the first display graph based on the hierarchical hypergraph module, the scheme update instruction is triggered. In response to the scheme update instruction, the knowledge hypergraph is updated according to the modified association relationship of the entity nodes and / or the content of the entity nodes, thereby updating the task processing scheme.

11. A task processing device, characterized in that, include: The knowledge hypergraph unit is used to generate a knowledge hypergraph based on the raw data and task processing prompts using a pre-set large model. The parameter generation unit is used to generate context parameters for rules based on the task processing requirements input by the user. The solution display unit is used to generate multiple task processing solutions based on the knowledge hypergraph and the context parameters using the preset large model, and to display the task processing solutions and the knowledge hypergraph through an interactive interface; A dynamic update unit is used to update the task processing scheme in response to a scheme update command, wherein the scheme update command is triggered when the original data is updated, the task processing requirements are modified, and / or the task processing scheme is modified, and the modification of the task processing requirements and the modification of the task processing scheme are implemented by the user based on the interactive interface.

12. An electronic device, characterized in that, include: One or more processors; Storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-10.

13. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, When the computer program or instructions are executed by a processor, they implement the method as described in any one of claims 1-10.

Images