Event solution plan generation method and device

Abstract

The invention discloses an event solution plan generation method and device, and belongs to the technical field of plan generation. The event solution plan generation method comprises the steps of obtaining event description information of a target event; based on the event description information, obtaining a plurality of target triples matched with target entities included in the event description information from a pre-constructed knowledge graph; the plurality of target triads comprise a plurality of historical entities corresponding to historical events and association relationships between target historical entities in the plurality of historical entities and other historical entities; the other historical entities are entities except the target historical entity in the plurality of historical entities; and reasoning the plurality of target triples to generate a target emergency plan of the target event. According to the event solution plan generation method, the target emergency plan of the target event is intelligently and specifically generated.

Description

Technical Field

[0001] The present application belongs to the technical field of plan generation, and in particular, relates to a method and device for generating an event resolution plan. Background Art

[0002] For various emergency events, the quick and effective generation of emergency management plans can greatly improve the efficiency of emergency management. In related technologies, emergency plans are mostly paper texts and unstructured electronic documents. The content of emergency plans is poorly related and poorly matched with the events to be solved, making it difficult to effectively assist emergency command decisions. Summary of the invention

[0003] The present application aims to solve at least one of the technical problems existing in the related art. To this end, the present application proposes a method and device for generating an event resolution plan, which intelligently and specifically generates a target emergency plan for a target event, improves the relevance and matching between the content in the target emergency plan and the target event, and thus improves the efficiency of assisting emergency command decision-making.

[0004] In a first aspect, the present application provides a method for generating an event resolution plan, the method comprising:

[0005] Get the event description information of the target event;

[0006] Based on the event description information, a plurality of target triples matching the target entity included in the event description information are obtained from a pre-constructed knowledge graph; the plurality of target triples include a plurality of historical entities corresponding to the historical event, and associations between the target historical entity and other historical entities in the plurality of historical entities; the other historical entities are entities other than the target historical entity in the plurality of historical entities;

[0007] Reasoning is performed on the multiple target triples to generate a target emergency plan for the target event.

[0008] According to the method for generating event resolution plans of the present application, by matching the target entities included in the acquired event description information with the triples in the pre-constructed knowledge graph, multiple target triples are obtained, and the multiple target triples are reasoned, and target emergency plans for target events are intelligently and targetedly generated, thereby improving the relevance and matching between the content in the target emergency plan and the target event, thereby improving the efficiency of auxiliary emergency command decision-making.

[0009] According to the method for generating an event resolution plan of the present application, the reasoning of the multiple target triples to generate a target emergency plan for the target event includes:

[0010] Based on the preset entity category and the association relationship between each of the target triples, reasoning is performed to obtain at least one emergency plan instruction; the multiple target triples include historical entities matching the target entity and entities corresponding to the preset entity category corresponding to the historical entity; and at least one of the target triples includes the emergency plan instruction;

[0011] Based on the at least one emergency plan instruction, the target emergency plan for the target event is generated.

[0012] According to the method for generating an event resolution plan of the present application, the preset entity categories include:

[0013] At least two of the categories corresponding to the event overview entity, the categories corresponding to the emergency disposal entity, the categories corresponding to the emergency response entity, and the categories corresponding to the safeguard measures entity.

[0014] According to the method for generating an event resolution plan of the present application, based on the event description information, a plurality of target triples matching the target entities included in the event description information are obtained from a pre-built knowledge graph, including:

[0015] Calculating the similarity between each historical entity in the knowledge graph and the target entity;

[0016] Based on each of the similarities, the multiple target triples matching the target entity are obtained from the knowledge graph.

[0017] According to the method for generating an event resolution plan of the present application, the step of obtaining the multiple target triples matching the target entity from the knowledge graph based on the similarities includes:

[0018] When the target similarity among the similarities is not less than a first similarity threshold, determining the historical entity corresponding to the target similarity as the historical entity matching the target entity;

[0019] Based on the historical entities, the multiple target triples are obtained.

[0020] According to the method for generating an event resolution plan of the present application, it is characterized in that the target entity included in the event description information is obtained through the following steps:

[0021] The event description information is input into a target recognition model to obtain the target entity included in the event description information output by the target recognition model; the target recognition model is obtained by integrating at least one of a bidirectional Transformer encoder, a bidirectional long short-term memory network, and a conditional random field.

[0022] According to the method for generating an event resolution plan of the present application, before obtaining the event description information of the target event, the method further includes:

[0023] An ontology model of a knowledge graph is constructed based on the acquired information of various historical events; the ontology model includes entity categories, relationships between entities, and attributes of entities that match the application scenario;

[0024] Based on the ontology model, extracting at least one historical entity matching the application scenario, the relationship between the historical entities, and the attributes of the historical entities from the multiple historical event information;

[0025] Based on the historical entities, the relationships between the historical entities, and the attributes of the historical entities, a knowledge graph corresponding to the application scenario is constructed.

[0026] According to the method for generating an event resolution plan of the present application, the knowledge graph corresponding to the application scenario is constructed based on the historical entities, the relationship between the historical entities, and the attributes of the historical entities, including:

[0027] Calculating the similarity between each of the historical entities;

[0028] When the similarity between a first historical entity in each of the historical entities and a second historical entity in each of the historical entities is greater than a second similarity threshold, fusing the first historical entity and the second historical entity to obtain a fused historical entity, a relationship between the fused historical entities, and an attribute of the fused historical entity;

[0029] The fused historical entities, the relationships among the fused historical entities, and the attributes of the fused historical entities are stored to construct the knowledge graph.

[0030] In a second aspect, the present application provides a device for generating an event resolution plan, the device comprising:

[0031] A first processing module, used to obtain event description information of a target event;

[0032] A second processing module is used to obtain, based on the event description information, a plurality of target triples matching the target entity included in the event description information from a pre-constructed knowledge graph; the plurality of target triples include a plurality of historical entities corresponding to the historical event, and associations between the target historical entity and other historical entities in the plurality of historical entities; the other historical entities are entities other than the target historical entity in the plurality of historical entities;

[0033] The third processing module is used to infer the multiple target triples to generate a target emergency plan for the target event.

[0034] According to the device for generating event resolution plans of the present application, by matching the target entities included in the acquired event description information with the triples in the pre-constructed knowledge graph, multiple target triples are obtained, and the multiple target triples are reasoned, so as to intelligently and specifically generate target emergency plans for target events, thereby improving the relevance and matching between the contents in the target emergency plans and the target events, and thus improving the efficiency of auxiliary emergency command decision-making.

[0035] In a third aspect, the present application provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for generating an event resolution plan as described in the first aspect above is implemented.

[0036] In a fourth aspect, the present application provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for generating an event resolution plan as described in the first aspect above.

[0037] In a fifth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the method for generating an event resolution plan as described in the first aspect above.

[0038] The above one or more technical solutions in the embodiments of the present application have at least one of the following technical effects:

[0039] By matching the target entities included in the acquired event description information with the triples in the pre-built knowledge graph, multiple target triples are obtained, and the multiple target triples are inferred to intelligently and specifically generate target emergency plans for target events, thereby improving the relevance and matching between the content in the target emergency plan and the target event, thereby improving the efficiency of auxiliary emergency command decision-making.

[0040] Furthermore, by presetting entity categories, the entity categories that need to be included in the content of the target emergency plan are determined, so that based on the required entity categories and the association between each target triple, each target triple is inferred to obtain multiple emergency plan instructions, thereby improving the efficiency and accuracy of reasoning on multiple target triples.

[0041] Furthermore, by calculating the similarity between each historical entity and the target entity in the knowledge graph, the qualitative information is quantified, the scientificity and efficiency of the matching process between historical entities and target entities are improved, and the multiple target triples that match the target entity are more accurate.

[0042] Furthermore, by analyzing the various historical event information obtained, the key entities in each historical event information, the relationships between entities, and the attributes of the entities are defined to construct an ontology model of the knowledge graph. Based on the defined ontology model, multiple historical entities in the historical event information, the relationships between each historical entity, and the attributes of each historical entity are extracted. Based on the multiple historical entities in the extracted historical event information, the relationships between each historical entity, and the attributes of each historical entity, a knowledge graph is constructed to effectively integrate various historical event information, make the relationships between each historical event information more intuitive, and improve the data quality and credibility of historical event information.

[0043] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0045] Figure 1 This is one of the flow charts of the method for generating an event resolution plan provided in an embodiment of the present application;

[0046] Figure 2 This is the second flow chart of the method for generating an event resolution plan provided in an embodiment of the present application;

[0047] Figure 3 This is the third flow chart of the method for generating an event resolution plan provided in an embodiment of the present application;

[0048] Figure 4 This is a fourth flow chart of a method for generating an event resolution plan provided in an embodiment of the present application;

[0049] Figure 5 This is the fifth flow chart of the method for generating an event resolution plan provided in the embodiment of the present application;

[0050] Figure 6 It is a structural schematic diagram of a device for generating an event resolution plan provided in an embodiment of the present application;

[0051] Figure 7 It is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0052] 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.

[0053] 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 data used in this way can be interchangeable under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described here, 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.

[0054] In combination with the accompanying drawings, the event resolution plan generation method, event resolution plan generation device, electronic device and readable storage medium provided in the embodiments of the present application are described in detail through specific embodiments and their application scenarios.

[0055] The method for generating an event resolution plan may be applied to a terminal, and may be specifically executed by hardware or software in the terminal.

[0056] The terminal includes but is not limited to portable communication devices such as mobile phones or tablet computers. It should also be understood that in some embodiments, the terminal may not be a portable communication device, but a desktop computer.

[0057] The terminal may include one or more other physical user interface devices such as a physical keyboard, mouse, and joystick.

[0058] The event resolution plan generation method provided in the embodiment of the present application may be executed by an electronic device or a functional module or functional entity in the electronic device that can implement the event resolution plan generation method. The electronic devices mentioned in the embodiment of the present application include but are not limited to mobile phones, tablet computers, computers, cameras, wearable devices, etc. The event resolution plan generation method provided in the embodiment of the present application is described below using an electronic device as an example of the execution subject.

[0059] like Figure 1 As shown, the method for generating the event resolution plan includes: step 110, step 120 and step 130.

[0060] Step 110: Obtain event description information of the target event;

[0061] In this step, the target event is the event to be resolved.

[0062] The target event may be an emergency event, for example, a flood disaster, a fire, an earthquake disaster or other emergency event, or a general event that does not occur suddenly.

[0063] The event description information is information describing the time, location, severity, etc. of the target event.

[0064] For example, the event description information may be that the water levels at multiple main control stations of River A exceed the warning water level.

[0065] The event description information may be voice information, text information, audio and video information, etc.

[0066] The specific content of the event description information can be customized by the user or determined based on actual conditions, and this application does not limit it.

[0067] In the actual execution process, the event description information of the target event can be obtained by collecting the voice, text input by the user or the audio and video files uploaded to understand the actual situation of the target event.

[0068] Step 120: Based on the event description information, a plurality of target triples matching the target entities included in the event description information are obtained from a pre-built knowledge graph;

[0069] In this step, the target entity is one or more entities included in the event description information.

[0070] It is understandable that event description information usually has no fixed requirements, and the description content is not exactly the same, and event description information may contain multiple synonyms.

[0071] In the actual execution process, the target entities included in the event description information can be obtained by semantically identifying the event description information, such as event danger, watershed location, river channel reservoir and other target entities.

[0072] The knowledge graph is a pre-constructed semantic network based on data such as historical emergency notifications, historical overall work plans, and historical command processes of various historical events.

[0073] The knowledge graph includes multiple triples.

[0074] A triplet includes entities and the relationships between entities.

[0075] Entities may include: event risk (such as flood, fire and earthquake), event time, event location, damage caused by the event, event severity level and event resolution measures.

[0076] The multiple target triples are triples in the knowledge graph that match the target entities included in the event description information.

[0077] Taking the event description information "Area A needs to carry out a level one flood prevention emergency response" as an example, the multiple target triples matched from the pre-built knowledge graph may be <flood, corresponding risk level, risk level>, <risk level, corresponding risk level, level one>, <flood, corresponding response plan, emergency plan> and <flood, corresponding response plan, emergency plan>, <emergency plan, required task instructions, water transfer>, etc.

[0078] The multiple target triples include multiple historical entities corresponding to the historical events, and association relationships between the target historical entity and other historical entities in the multiple historical entities.

[0079] Historical entities are entities obtained by extracting various historical events.

[0080] The target historical entity is any historical entity among the multiple historical entities.

[0081] The other historical entities are entities other than the target historical entity among the multiple historical entities.

[0082] During the actual execution process, the event description information can be semantically recognized through deep learning models and machine learning models to obtain the target entity included in the event description information. Based on the target entity, multiple target triples matching the target entity are matched from the pre-built knowledge graph.

[0083] Step 130: Infer multiple target triples to generate a target emergency plan for the target event.

[0084] In this step, the target emergency plan is a pre-made solution to deal with the target event.

[0085] In the actual execution process, after obtaining multiple target triples matching the target entity, the multiple target triples can be inferred, that is, new unknown triples can be derived through the known multiple target triples.

[0086] In the actual execution process, multiple target triples can be inferred through one or more of the rule-based reasoning method, the logic-based reasoning method and the representation learning-based reasoning method to obtain one or more unknown triples, thereby generating a target emergency plan for the target event based on the obtained unknown triples.

[0087] Taking the flood in area A as an example, the multiple target triples obtained may be <flood, corresponding response plan, emergency plan><emergency plan, required task instructions, water transfer>. Through the multiple target triples obtained, we can get <flood, required task instructions, water transfer>.

[0088] It is understandable that by reasoning on multiple known target triples, multiple unknown triples can be obtained, so that based on the multiple unknown triples, the way to deal with the target event can be determined to obtain the target emergency plan.

[0089] During the research and development process, the inventors found that the quick and effective generation of emergency management plans for various emergency events can greatly improve the efficiency of emergency management. In the related art, flood control emergency plans are mostly paper texts and unstructured electronic documents. The content of the emergency plans is poorly related and poorly matched with the events to be solved, making it difficult to effectively assist emergency command decision-making.

[0090] The present application obtains event description information related to the target event, understands the characteristics of the target event, and obtains multiple target triples that match the target entity included in the event description information from a pre-constructed knowledge graph, obtains multiple historical entities similar to the target entity in the historical event and the association relationship between the historical entities, thereby performing reasoning based on the multiple historical entities and the association relationship between the historical entities, and intelligently generates a target emergency plan required to resolve the target event. Based on the characteristics of the target event, the target emergency plan is generated in a targeted manner, which can improve the relevance and matching of the content in the target emergency plan with the target event, thereby improving the efficiency of auxiliary emergency command decision-making.

[0091] According to the method for generating an event resolution plan provided in an embodiment of the present application, a plurality of target triples are obtained by matching the target entities included in the acquired event description information with the triples in the pre-constructed knowledge graph, and the plurality of target triples are reasoned, so as to intelligently and specifically generate a target emergency plan for the target event, thereby improving the relevance and matching between the content in the target emergency plan and the target event, and thus improving the efficiency of auxiliary emergency command decision-making.

[0092] In some embodiments, the target entity included in the event description information may be obtained by the following steps:

[0093] The event description information is input into the target recognition model, and the target entity included in the event description information output by the target recognition model is obtained.

[0094] In this embodiment, the target recognition model is obtained by integrating at least one of a bidirectional Transformer encoder, a bidirectional long short-term memory network, and a conditional random field.

[0095] The bidirectional Transformer encoder is trained on Google's large-scale general Chinese corpus and has strong ability to acquire general semantic information.

[0096] The bidirectional LSTM network combines the outputs of the forward and backward LSTM networks to capture bidirectional dependencies in the sequence.

[0097] Conditional random fields are conditional probability distribution models of another set of output random variables Y given a set of input random variables X.

[0098] The target recognition model is a recognition model obtained by integrating one or more of a bidirectional Transformer encoder, a bidirectional long short-term memory network, and a conditional random field.

[0099] In some embodiments, the fusion of bidirectional long short-term memory network and conditional random field can obtain a classic named entity recognition model.

[0100] The emission score (i.e., label vector) is generated through a bidirectional long short-term memory network. Based on the emission score, the optimal label sequence is decoded through the conditional random field. This label sequence can be used as the output of the model, that is, to find the path with the highest probability and the best effect among all possible paths.

[0101] In some embodiments, a bidirectional Transformer encoder, a bidirectional long short-term memory network, and a conditional random field may be fused to obtain a target recognition model.

[0102] The target entity included in the event description information is identified through the target recognition model.

[0103] In the actual implementation process, Figure 2 As shown in the figure, the event description information "Xiao Ni River dike breach" is input into the target recognition model. The target recognition model converts the event description information into a machine-recognizable digital code list through the encoding layer and passes it into the bidirectional Transformer encoder.

[0104] The bidirectional Transformer encoder extracts common semantic features from event description information to generate word vectors. The bidirectional long short-term memory network is used to make the word vectors carry the contextual semantic information of the event description information. The fully connected layer is used to convert the word vectors into flood control emergency entity labels. Finally, the conditional random field is used to correct the entity labels and output the optimal label sequence.

[0105] During the research and development process, the inventors discovered that in the related technologies, the methods for entity recognition are relatively simple and crude, often applicable to small-scale and single-form data, and difficult to deal with the problem of multiple synonyms.

[0106] The present application performs entity recognition by inputting event description information into a target recognition model obtained by fusing one or more of a bidirectional Transformer encoder, a bidirectional long short-term memory network, and a conditional random field. By combining the advantages of a bidirectional Transformer encoder, a bidirectional long short-term memory network, and a conditional random field, the semantic acquisition capability of the target recognition model is improved, and better entity recognition results can be obtained, thereby improving the accuracy and efficiency of target entity recognition in event description information.

[0107] According to the method for generating event resolution plans provided in the embodiments of the present application, by fusing one or more of a bidirectional Transformer encoder, a bidirectional long short-term memory network, and a conditional random field, and combining the advantages of multiple models, the target entities in the event description information can be intelligently identified, and large-scale and complex event description information can be effectively responded to. It can accurately process information with multiple synonyms, and improve the accuracy and efficiency of identifying target entities in the event description information.

[0108] In some embodiments, step 120 may further include:

[0109] Calculate the similarity between each historical entity and the target entity in the knowledge graph;

[0110] Based on each similarity, multiple target triples matching the target entity are obtained from the knowledge graph.

[0111] In this embodiment, the similarity is used to measure the similarity between the historical entity and the target entity.

[0112] It is understandable that the target entity identified in the event description information may be different from the entity in the knowledge graph, and there may be abbreviations and multiple synonyms. By comparing the similarity between the historical entity and the target entity, multiple target triples corresponding to historical entities similar to the target entity included in the event description information can be matched from the pre-built knowledge graph.

[0113] In the actual execution process, the similarity between the historical entity and the target entity can be calculated based on the statistical method of the context vector space.

[0114] Based on the statistical method of context vector space, we can use Google's word vectorization tool (Word toVector, Word2Vec) to map the words in the event description information to space vectors, and measure the similarity between the historical entity and the target entity by calculating the vector distance between the two entities. That is, through the entity semantic alignment technology, we can match the target event with the historical events in the knowledge graph.

[0115] According to the method for generating event resolution plans provided in the embodiments of the present application, by calculating the similarity between each historical entity and the target entity in the knowledge graph, the qualitative information is quantified, the scientificity and efficiency of the matching process between historical entities and target entities are improved, and the multiple target triples obtained that match the target entity are more accurate.

[0116] In some embodiments, based on each similarity, obtaining multiple target triples matching the target entity from the knowledge graph may also include:

[0117] When the target similarity among the similarities is not less than the first similarity threshold, determining the historical entity corresponding to the target similarity as the historical entity matching the target entity;

[0118] Based on historical entities, multiple target triplets are obtained.

[0119] In this embodiment, the target similarity is any similarity among the multiple similarities obtained by calculation.

[0120] The first similarity threshold is a preset value for determining whether the historical entity is similar to the target entity.

[0121] The specific value of the first similarity threshold can be customized by the user or determined based on actual conditions, and is not limited in this application.

[0122] During the actual execution process, the correlation between each similarity and the first similarity threshold can be compared one by one. When the target similarity among each similarity is not less than the first similarity threshold, it is approximately considered that the historical entity corresponding to the target similarity has a high similarity with the target entity or the historical entity is equivalent to the target entity, and the historical entity corresponding to the target similarity is determined as the historical entity that matches the target entity.

[0123] After obtaining the historical entity that matches the target entity, the association relationship between the historical entity and other historical entities in the knowledge graph is extracted to obtain the target triple.

[0124] It is understandable that the calculated similarity may include multiple target similarities that are not less than the first similarity threshold, thereby obtaining multiple historical entities matching the target entity, and obtaining multiple target triples based on the multiple historical entities.

[0125] According to the method for generating an event resolution plan provided in an embodiment of the present application, a first similarity threshold is set to screen historical entities with higher similarities among various similarities, thereby obtaining historical entities with higher similarity to the target entities included in the event description information, thereby improving the correlation between multiple target triples obtained based on the historical entities and the target events, thereby improving the accuracy of the subsequently generated target emergency plans.

[0126] In some embodiments, step 130 may further include:

[0127] Based on the association relationship between the preset entity categories and each target triple, reasoning is performed to obtain at least one emergency plan instruction; the multiple target triples include historical entities matching the target entity and entities corresponding to the preset entity categories corresponding to the historical entities; and at least one target triple includes an emergency plan instruction;

[0128] Based on at least one emergency plan instruction, a target emergency plan for a target event is generated.

[0129] In this embodiment, the preset entity category is the entity category required for generating the target emergency plan.

[0130] The specific categories included in the preset entity categories can be determined based on user customization or based on actual conditions, and this application does not limit this.

[0131] In some embodiments, the preset entity categories may include:

[0132] At least two of the categories corresponding to the event overview entity, the categories corresponding to the emergency disposal entity, the categories corresponding to the emergency response entity, and the categories corresponding to the safeguard measures entity.

[0133] In this embodiment, the event overview entity is an entity that describes the event situation.

[0134] The event overview entity may include: event name, event type, location, impact range, urgency, and related dangerous events, etc.

[0135] The emergency response entity is the entity corresponding to the response method for the target event.

[0136] Emergency response entities may include: information reporting, advance response, and information release, etc.

[0137] The emergency response entity is the entity corresponding to the response target event.

[0138] Emergency response entities may include: emergency response conditions at all levels and response actions, etc.

[0139] The safeguard entity is the entity corresponding to the safeguard required to resolve the target incident.

[0140] The entities of the guarantee measures include: emergency team guarantee, equipment and material guarantee, and medical and health care guarantee, etc.

[0141] In this embodiment, the user can select the required entity category from the category corresponding to the event overview entity, the category corresponding to the emergency disposal entity, the category corresponding to the emergency response entity, and the category corresponding to the safeguard measure entity based on actual usage requirements.

[0142] For example, when a user needs to obtain a way to respond to a target event based on the actual situation of the target event, the preset entity categories may include an event overview entity and an emergency response entity.

[0143] For another example, when a user needs to obtain a method for responding to a target event and materials required for the target event based on the actual situation of the target event, the preset entity categories may include an event overview entity, an emergency response entity, and a safeguard measure entity.

[0144] In some embodiments, an emergency plan template may be constructed based on preset entity categories.

[0145] In this embodiment, the emergency plan template is a template including entity categories and format settings that the target emergency plan needs to include.

[0146] According to the method for generating an event resolution plan provided in an embodiment of the present application, by determining preset entity categories including categories corresponding to event overview entities, categories corresponding to emergency disposal entities, categories corresponding to emergency response entities, and categories corresponding to safeguard measures entities, the information required for generating a target emergency plan can be effectively determined, and a standardized target emergency plan can be generated, thereby improving the pertinence of the generated target emergency plan and the user experience.

[0147] In the actual execution process, after obtaining multiple target triples, the multiple target triples can be inferred based on preset entity categories through graph retrieval and reasoning technology to obtain one or more emergency plan instructions, thereby generating a target emergency plan based on the one or more emergency plan instructions obtained by reasoning.

[0148] Graph retrieval reasoning technology includes establishing rules and making inferences.

[0149] In the actual execution process, users can pre-establish inference rules, such as one-hop inference rules, two-hop inference rules, and multi-hop inference rules.

[0150] Among them, the one-hop reasoning rule is to infer unknown entities through known entities and the relationships between entities.

[0151] For example, if the one-hop inference rule is South Street Warehouse > Number of Tents > X, and it is written in Structured Query Language (SQL), the number of all tents in the South Street Warehouse can be queried.

[0152] The two-hop reasoning rule is to infer an unknown triple through two target triples and the association between the two target triples.

[0153] For example, the two-hop reasoning rule is <emergency, corresponding emergency plan, emergency plan><emergency plan, including work arrangements, instructions> and is deduced to <emergency, required task instructions, instructions>. By writing an SQL query statement, it is possible to output an unknown triple from at least two target triples existing in the graph.

[0154] That is, two target triples <extremely heavy rain events, corresponding emergency plans, flood control level I work deployment Y><flood control level I work deployment Y, including work arrangements, launching joint duty mechanism publicity and reporting, etc.>, lead to an unknown triplet <extremely heavy rain events, required task instructions, launching joint duty mechanism publicity and reporting, etc.>.

[0155] The multi-hop reasoning rule is to infer an unknown triple through multiple target triples and the associations between multiple target triples.

[0156] Of course, each target triple may also be inferred by any other feasible method, such as chain reasoning and path reasoning, etc., which is not limited in this application.

[0157] In the actual implementation process, the entity categories that need to be included in the generated target emergency plan can be determined based on the preset entity categories, and multiple target triples can be inferred through graph retrieval and reasoning technology to obtain multiple emergency plan instructions, thereby generating a target emergency plan based on multiple emergency plan instructions.

[0158] In some embodiments, each emergency plan instruction obtained by reasoning can be filled into an emergency plan template to generate a target emergency plan.

[0159] like Figure 3 As shown, after obtaining the event description information of the target event, entity recognition is performed on the event description information to obtain the target entity, such as an emergency, a location, a river reservoir, etc. Then, the target entity is matched with the historical entity in the pre-constructed knowledge graph, that is, through the entity alignment technology, multiple target triples are obtained, and based on the preset entity categories, the multiple target triples are inferred through the graph retrieval reasoning technology to generate a target emergency plan.

[0160] According to the method for generating an event resolution plan provided in an embodiment of the present application, by presetting entity categories, the entity categories that need to be included in the content of the target emergency plan are determined, so that based on the required entity categories and the association relationship between each target triplet, each target triplet is inferred to obtain multiple emergency plan instructions, thereby improving the efficiency and accuracy of reasoning on multiple target triples.

[0161] In some embodiments, before step 110, the method may further include:

[0162] Construct an ontology model of the knowledge graph based on the acquired information of various historical events;

[0163] Based on the ontology model, extract at least one historical entity matching the application scenario, the relationship between the historical entities, and the attributes of the historical entities from the various historical event information;

[0164] Based on each historical entity, the relationship between each historical entity, and the attributes of each historical entity, a knowledge graph corresponding to the application scenario is constructed.

[0165] In this embodiment, the ontology model is a model obtained by formally describing various historical event information based on actual application scenarios.

[0166] The ontology model includes entity categories that match the application scenario, the relationships between entities, and the attributes of the entities.

[0167] The application scenario is the scenario in which the events that need to be processed occur during the actual execution process.

[0168] Application scenarios may include: flood prevention scenarios, fire scenarios, earthquake scenarios, and typhoon scenarios, etc.

[0169] Historical event information is information related to events that occurred before the current collection time.

[0170] Historical event information may include: emergency notification, overall emergency work plan, emergency command process and other information.

[0171] In the actual implementation process, after obtaining various historical event information, various historical event information is used as the data source, and the emergency plan ontology and the relationship between ontologies are defined in a top-down and bottom-up manner to construct an ontology model of the knowledge graph.

[0172] Taking the flood control emergency plan as an example, the knowledge graph ontology model of the flood control emergency plan is as follows: Figure 4 As shown in the figure, the knowledge graph ontology model of flood emergency plan includes entities such as emergencies, control measures, river reservoirs, and information reporting, as well as information such as the relationship between emergencies and control measures.

[0173] In the actual execution process, after obtaining the ontology model, multiple historical entities, the relationships between historical entities, and the attributes of each historical entity can be extracted from the historical event information based on the entities, the relationships between entities, and the attributes of the entities defined by the ontology model.

[0174] It is understandable that most of the historical event information is unstructured and semi-structured documents and flow charts, which makes it difficult to construct effective learning samples.

[0175] In the actual implementation process, entities in various historical event information can be extracted based on the "keyword + short text" method to construct learning samples.

[0176] As shown in Table 1, taking the historical event information of flood control events as an example, the content extracted from “keywords + short text” is explained.

[0177] Table 1

[0178]

[0179] The attributes of a historical entity are elements or information that describe the characteristics or status of the historical entity.

[0180] The attributes of an emergency event may include risk factors and location, etc.

[0181] In some embodiments, each historical entity, the relationship between each historical entity, and the attributes of each historical entity may be stored in a graph database.

[0182] In this embodiment, the graph database has the characteristics of higher traversal efficiency and stronger relationship expression ability.

[0183] It should be noted that generating target emergency plans has high requirements for the expression relationship.

[0184] In the actual implementation process, the Neo4j native graph database with higher entity relationship query efficiency can be used to store each historical entity, the relationship between each historical entity, and the attributes of each historical entity.

[0185] In the actual execution process, after extracting at least one historical entity, the relationship between historical entities, and the attributes of each historical entity from a variety of historical event information, the historical entities, the relationship between historical entities, and the attributes of each historical entity can be organized into a triple format <entity 1, relationship, entity 2>, such as <flood disaster, post-disaster impact, farmland destruction>, and the multiple triples obtained by organization are stored to construct a knowledge graph.

[0186] According to the method for generating event resolution plans provided in the embodiments of the present application, by analyzing the various historical event information obtained, the key entities in each historical event information, the relationships between entities, and the attributes of the entities are defined to construct an ontology model of the knowledge graph. Based on the defined ontology model, multiple historical entities in the historical event information, the relationships between the historical entities, and the attributes of the historical entities are extracted. Based on the multiple historical entities in the extracted historical event information, the relationships between the historical entities, and the attributes of the historical entities, a knowledge graph is constructed to effectively integrate the various historical event information, make the relationships between the historical event information more intuitive, and improve the data quality and credibility of the historical event information.

[0187] In some embodiments, based on the historical entities, the relationships between the historical entities, and the attributes of the historical entities, constructing a knowledge graph corresponding to the application scenario may also include:

[0188] Calculate the similarity between historical entities;

[0189] When the similarity between a first historical entity in each historical entity and a second historical entity in each historical entity is greater than a second similarity threshold, the first historical entity and the second historical entity are fused to obtain a fused historical entity, a relationship between the fused historical entities, and an attribute of the fused historical entity;

[0190] The fused historical entities, the relationships among the fused historical entities, and the attributes of the fused historical entities are stored to build a knowledge graph.

[0191] In this embodiment, the first historical entity is any historical entity among a plurality of historical entities.

[0192] The second historical entity is any historical entity among the plurality of historical entities except the first historical entity.

[0193] The second similarity threshold is a preset value for determining that two historical entities are similar.

[0194] The specific value of the second similarity threshold can be customized by the user or determined based on actual conditions, and is not limited in this application.

[0195] The fused historical entity is an entity that retains the first historical entity information and the second historical entity information.

[0196] The fused historical entity may be the first historical entity, the second historical entity, or a new historical entity representing common features of the first historical entity and the second historical entity.

[0197] In the actual implementation process, the similarity between historical entities can be calculated through statistical methods of context vector space and cosine similarity.

[0198] Of course, the similarity between entities may also be calculated in any other feasible manner, which is not limited in this application.

[0199] When the similarity between the first historical entity and the second historical entity in the calculated similarity is greater than the second similarity threshold, the first historical entity and the second historical entity are approximately considered to be the same, the first historical entity and the second historical entity are fused to obtain a fused historical entity, the relationship between the first historical entity and other historical entities and the relationship between the second historical entity and other historical entities are fused, and the attributes of the first historical entity and the attributes of the second historical entity are fused; the fused historical entity, the relationship between the fused historical entities and the attributes of the fused historical entity are obtained, the fused historical entity, the relationship between the fused historical entities and the attributes of the fused historical entity are stored, and a knowledge graph is constructed.

[0200] It is understandable that there may be more than two similar historical entities among the historical entities, and similar historical entities can be used as candidate historical entities and fused to obtain a fused historical entity.

[0201] Taking "the water level at multiple main control stations of River A exceeds the warning level" as an example, the entity of the historical event extracted is "exceeding the warning level". After the context entity reference, it can be expanded to "the water level at the control station exceeds the warning level". A group of entities "multiple warning water level warnings, river warning water level warnings, and embankment breaches, etc." are generated through candidate entities. Finally, by sorting the candidate entities, the candidate entities equivalent to the historical event entity "multiple warning water level warnings" in the knowledge graph are fused to obtain the fused historical entity.

[0202] like Figure 5 As shown, knowledge modeling can be performed based on actual application scenarios, and entities, relationships between entities, and attributes of entities related to the application scenarios can be defined to obtain an ontology model. Then, based on the entities, relationships between entities, and attributes of entities included in the ontology model, entities, relationships between entities, and attributes of entities in historical event information can be extracted, and disambiguation and reference resolution operations can be performed on the extracted entities, that is, knowledge fusion can be performed, and the fused knowledge can be stored in a graph database through a resource description framework to obtain a knowledge graph.

[0203] According to the method for generating event resolution plans provided in the embodiments of the present application, by calculating the similarity between historical entities, any two historical entities whose similarity is greater than a second similarity threshold are fused, and each entity is disambiguated and reference resolved to improve the quality of the entity and the consistency of the entity, thereby improving the quality and consistency of the fused historical entities, thereby improving the quality, consistency and completeness of the constructed knowledge graph.

[0204] The event solution plan generation method provided in the embodiment of the present application can be executed by a device for generating event solution plans. In the embodiment of the present application, the device for generating event solution plans executing the event solution plan generation method is used as an example to illustrate the device for generating event solution plans provided in the embodiment of the present application.

[0205] The embodiment of the present application also provides a device for generating an event resolution plan.

[0206] like Figure 6 As shown, the device for generating an event resolution plan includes: a first processing module 610 , a second processing module 620 and a third processing module 630 .

[0207] A first processing module 610 is used to obtain event description information of a target event;

[0208] The second processing module 620 is used to obtain, based on the event description information, a plurality of target triples matching the target entity included in the event description information from a pre-built knowledge graph; the plurality of target triples include a plurality of historical entities corresponding to the historical event, and an association relationship between the target historical entity and other historical entities in the plurality of historical entities; the other historical entities are entities other than the target historical entity in the plurality of historical entities;

[0209] The third processing module 630 is used to infer multiple target triples to generate a target emergency plan for the target event.

[0210] According to the device for generating event resolution plans provided in the embodiment of the present application, multiple target triples are obtained by matching the target entities included in the acquired event description information with the triples in the pre-constructed knowledge graph, and the multiple target triples are reasoned to intelligently and specifically generate target emergency plans for target events, thereby improving the relevance and matching between the content in the target emergency plan and the target event, thereby improving the efficiency of auxiliary emergency command decision-making.

[0211] In some embodiments, the third processing module 630 may also be used to:

[0212] Based on the association relationship between the preset entity categories and each target triple, reasoning is performed to obtain at least one emergency plan instruction; the multiple target triples include historical entities matching the target entity and entities corresponding to the preset entity categories corresponding to the historical entities; and at least one target triple includes an emergency plan instruction;

[0213] Based on at least one emergency plan instruction, a target emergency plan for a target event is generated.

[0214] In some embodiments, the second processing module 620 may also be used to:

[0215] Calculate the similarity between each historical entity and the target entity in the knowledge graph;

[0216] Based on each similarity, multiple target triples matching the target entity are obtained from the knowledge graph.

[0217] In some embodiments, the second processing module 620 may also be used to:

[0218] When the target similarity among the similarities is not less than the first similarity threshold, determining the historical entity corresponding to the target similarity as the historical entity matching the target entity;

[0219] Based on historical entities, multiple target triplets are obtained.

[0220] In some embodiments, the apparatus may further include a fourth processing module, configured to:

[0221] The event description information is input into the target recognition model to obtain the target entity included in the event description information output by the target recognition model; the target recognition model is obtained by integrating at least one of a bidirectional Transformer encoder, a bidirectional long short-term memory network and a conditional random field.

[0222] In some embodiments, the apparatus may further include a fifth processing module, configured to:

[0223] An ontology model of the knowledge graph is constructed based on the acquired information of various historical events; the ontology model includes entity categories matching the application scenario, the relationships between entities, and the attributes of entities;

[0224] Based on the ontology model, extract at least one historical entity matching the application scenario, the relationship between the historical entities, and the attributes of the historical entities from the various historical event information;

[0225] Based on each historical entity, the relationship between each historical entity, and the attributes of each historical entity, a knowledge graph corresponding to the application scenario is constructed.

[0226] In some embodiments, the apparatus may further include a sixth processing module, configured to:

[0227] Calculate the similarity between historical entities;

[0228] When the similarity between a first historical entity in each historical entity and a second historical entity in each historical entity is greater than a second similarity threshold, the first historical entity and the second historical entity are fused to obtain a fused historical entity, a relationship between the fused historical entities, and an attribute of the fused historical entity;

[0229] The fused historical entities, the relationships among the fused historical entities, and the attributes of the fused historical entities are stored to build a knowledge graph.

[0230] The generating device of the event solution plan 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 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.

[0231] The device for generating an event resolution plan 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.

[0232] The device for generating an event solution plan provided in the embodiment of the present application can achieve Figures 1 to 5 To avoid repetition, the various processes implemented by the method embodiment are not described here.

[0233] In some embodiments, Figure 7 As shown, an embodiment of the present application also provides an electronic device 700, including a processor 701, a memory 702, and a computer program stored in the memory 702 and executable on the processor 701. When the program is executed by the processor 701, each process of the embodiment of the method for generating a method for a proposed event resolution is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0234] 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.

[0235] An embodiment of the present application also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the various processes of the above-mentioned event resolution plan generation method embodiment are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0236] 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.

[0237] An embodiment of the present application also provides a computer program product, including a computer program, which implements the above-mentioned method for generating an event resolution plan when executed by a processor.

[0238] 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.

[0239] 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 event resolution plan generation method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0240] 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.

[0241] 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.

[0242] 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 relevant technology, 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.

[0243] 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.

[0244] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0245] Although the embodiments of the present application have been shown and described, those skilled in the art will appreciate that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present application, and that the scope of the present application is defined by the claims and their equivalents.

Claims

1. A method for generating an event resolution plan, characterized in that: include: Get the event description information of the target event; Based on the event description information, a plurality of target triples matching the target entity included in the event description information are obtained from a pre-constructed knowledge graph; the plurality of target triples include a plurality of historical entities corresponding to the historical event, and associations between the target historical entity and other historical entities in the plurality of historical entities; the other historical entities are entities other than the target historical entity in the plurality of historical entities; Reasoning is performed on the multiple target triples to generate a target emergency plan for the target event.

2. The method for generating an event resolution plan according to claim 1, characterized in that: The reasoning of the multiple target triples to generate a target emergency plan for the target event includes: Based on the preset entity category and the association relationship between each of the target triples, reasoning is performed to obtain at least one emergency plan instruction; the multiple target triples include historical entities matching the target entity and entities corresponding to the preset entity category corresponding to the historical entity; and at least one of the target triples includes the emergency plan instruction; Based on the at least one emergency plan instruction, the target emergency plan for the target event is generated.

3. The method for generating an event resolution plan according to claim 2, characterized in that: The preset entity categories include: At least two of the categories corresponding to the event overview entity, the categories corresponding to the emergency disposal entity, the categories corresponding to the emergency response entity, and the categories corresponding to the safeguard measures entity.

4. The method for generating an event resolution plan according to any one of claims 1 to 3, characterized in that: The step of acquiring, based on the event description information, a plurality of target triples matching the target entities included in the event description information from a pre-built knowledge graph includes: Calculating the similarity between each historical entity in the knowledge graph and the target entity; Based on each of the similarities, the multiple target triples matching the target entity are obtained from the knowledge graph.

5. The method for generating an event resolution plan according to claim 4, characterized in that: The acquiring, based on each of the similarities, the plurality of target triples matching the target entity from the knowledge graph comprises: When the target similarity among the similarities is not less than a first similarity threshold, determining the historical entity corresponding to the target similarity as the historical entity matching the target entity; Based on the historical entities, the multiple target triples are obtained.

6. The method for generating an event resolution plan according to any one of claims 1 to 3, characterized in that: The target entity included in the event description information is obtained by the following steps: The event description information is input into a target recognition model to obtain the target entity included in the event description information output by the target recognition model; the target recognition model is obtained by integrating at least one of a bidirectional Transformer encoder, a bidirectional long short-term memory network, and a conditional random field.

7. The method for generating an event resolution plan according to any one of claims 1 to 3, characterized in that: Before acquiring the event description information of the target event, the method further includes: An ontology model of a knowledge graph is constructed based on the acquired information of various historical events; the ontology model includes entity categories, relationships between entities, and attributes of entities that match the application scenario; Based on the ontology model, extracting at least one historical entity matching the application scenario, the relationship between the historical entities, and the attributes of the historical entities from the multiple historical event information; Based on the historical entities, the relationships between the historical entities, and the attributes of the historical entities, a knowledge graph corresponding to the application scenario is constructed.

8. The method for generating an event resolution plan according to claim 7, characterized in that: The constructing of a knowledge graph corresponding to the application scenario based on the historical entities, the relationships between the historical entities, and the attributes of the historical entities includes: Calculating the similarity between each of the historical entities; When the similarity between a first historical entity in each of the historical entities and a second historical entity in each of the historical entities is greater than a second similarity threshold, fusing the first historical entity and the second historical entity to obtain a fused historical entity, a relationship between the fused historical entities, and an attribute of the fused historical entity; The fused historical entities, the relationships among the fused historical entities, and the attributes of the fused historical entities are stored to construct the knowledge graph.

9. A device for generating an event resolution plan, characterized in that: include: A first processing module, used to obtain event description information of a target event; A second processing module is used to obtain, based on the event description information, a plurality of target triples matching the target entity included in the event description information from a pre-constructed knowledge graph; the plurality of target triples include a plurality of historical entities corresponding to the historical event, and associations between the target historical entity and other historical entities in the plurality of historical entities; the other historical entities are entities other than the target historical entity in the plurality of historical entities; The third processing module is used to infer the multiple target triples to generate a target emergency plan for the target event.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, it implements the method for generating an event resolution plan as described in any one of claims 1-8.

Images