Flight safety special situation disposal method, device, equipment and medium
By constructing a special situation knowledge vector library and using large language models to generate and dispose solutions, the bias problem of relying on empirical knowledge in the existing technology is solved, the accuracy and accuracy of flight safety special situation treatment is improved, and the special characteristics of complex and multi-round dialogue are adapted to Situation handling.
Patent Information
- Application Number
- CN202411084945.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-08
- Publication Date
- 2024-11-22
AI Technical Summary
The existing technology over-reliance on the commander/pilot's experience and knowledge in handling flight safety situations is susceptible to cognitive bias, and the online Q&A method based on the knowledge base has shortcomings in understanding complex situations and handling multiple rounds of dialogues, which is difficult to Handle ambiguous or complex user input.
By constructing a special situation knowledge vector library, retrieve associated special situation knowledge data based on target special situation problems, generate prompt words and use large language models to generate a disposal plan, determine the target disposal plan based on the average similarity of associated special situation knowledge data, and improve the disposal plan accuracy and accuracy.
It improves the accuracy and accuracy of flight safety special situation handling, reduces the dependence on commander/pilot experience, and can more effectively deal with special situation problems in complex and multi-round dialogue.
Smart Images

Figure CN119003726A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of computer technology and related technical fields, and in particular to a method, device, equipment and medium suitable for handling special flight safety situations. Background Art
[0002] Handling special situations in the flight field is one of the important means to ensure flight safety, improve flight safety emergency response capabilities, and optimize resource allocation.
[0003] In the prior art, the corresponding special situations are mainly handled based on the experience and knowledge accumulated by the commander / pilot, and the online question-and-answer method based on the knowledge base provides auxiliary solutions. The special situation handling method based on experience and knowledge will over-rely on the knowledge and experience of the commander or pilot, resulting in the neglect of special situations. At the same time, it is easily affected by cognitive biases, such as confirmation bias (the commander / pilot only pays attention to information consistent with his or her existing knowledge) and anchoring effect (over-reliance on the initially obtained information and inability to fully control other information in the process of utilizing the special situation). These biases may affect the judgment and decision-making of special situations. Due to the relatively complex problems and knowledge related to the handling of special situations, the online question-and-answer method based on the knowledge base has deficiencies in understanding the context and implicit meaning of complex special situation problems, especially when dealing with multiple rounds of dialogue and problems that require comprehensive information on historical special situation cases, special situation rules, etc. At the same time, its natural language processing capabilities are limited, and it cannot fully understand and generate corresponding special situation problems and handling plans, and it is difficult to handle ambiguous, unclear or complex user input. Summary of the invention
[0004] The embodiments described herein provide a method, device, equipment and medium for handling special flight safety situations to solve the problems existing in the prior art.
[0005] In a first aspect, according to the content of the present disclosure, a method for handling a flight safety special situation is provided, comprising:
[0006] According to the target special situation problem, a first target number of associated special situation knowledge data associated with the target special situation problem is retrieved from a special situation knowledge vector library, wherein the special situation knowledge vector library includes a plurality of special situation knowledge data, and the special situation knowledge data includes accident information, disposal plans and associated documents;
[0007] Generate a prompt word for the target special situation question according to the target special situation question, the first target number of associated special situation knowledge data and a prompt word template;
[0008] Generate a solution to the target special situation problem according to the prompt word of the target special situation problem;
[0009] A target handling solution for the target special situation problem is determined based on an average similarity between the handling solution for the target special situation problem and the handling solutions included in each of the associated special situation knowledge data.
[0010] In some embodiments of the present disclosure, the step of retrieving a first target number of associated special situation knowledge data associated with the target special situation problem from a special situation knowledge vector library according to the target special situation problem includes:
[0011] Performing vectorization processing on the target special situation problem to obtain a target special situation problem vector;
[0012] Respectively performing similarity matching between the target special situation problem vector and the accident information included in each of the special situation knowledge data to obtain a similarity value;
[0013] After the similarity values of the target special situation problem vector and the accident information included in each of the special situation knowledge data are sorted in reverse order, the special situation knowledge data of the first target quantity are selected to form the associated special situation knowledge data.
[0014] In some embodiments of the present disclosure, generating a prompt word for the target special situation problem according to the target special situation problem, the first target number of associated special situation knowledge data and a prompt word template includes:
[0015] Determining a special situation handling context input for the target special situation problem according to the associated documents included in the first target number of associated special situation knowledge data;
[0016] The target special situation problem and the special situation handling context are input into a prompt word template to generate a prompt word for the target special situation problem.
[0017] In some embodiments of the present disclosure, generating a solution to the target special situation problem according to the prompt word of the target special situation problem includes:
[0018] The prompt words of the target special situation problem are input into the large language model, and the prompt words of the target special situation problem are analyzed and identified based on the large language model to obtain a solution to the target special situation problem.
[0019] In some embodiments of the present disclosure, determining a target solution for the target special situation problem according to an average similarity between the solution for the target special situation problem and the solution included in each of the associated special situation knowledge data includes:
[0020] Determining in sequence the similarity between the solution for the target special situation problem and the solution included in each of the associated special situation knowledge data;
[0021] The similarities between the treatment plan of the target special situation problem and the treatment plan corresponding to each of the associated special situation knowledge data are summed and averaged to obtain the average similarity between the treatment plan of the target special situation problem and the treatment plan included in each of the associated special situation knowledge data;
[0022] The target handling solution for the target special situation problem is determined based on the relationship between the average similarity between the handling solution for the target special situation problem and the handling solutions included in each of the associated special situation knowledge data and a preset similarity threshold.
[0023] In some embodiments of the present disclosure, determining a target solution for the target special situation problem according to an average similarity between the solution for the target special situation problem and the solution included in each of the associated special situation knowledge data includes:
[0024] When the average similarity between the treatment plan of the target special situation problem and the treatment plans included in each of the associated special situation knowledge data meets the preset similarity threshold, the treatment plan generated according to the prompt word of the target special situation problem is the target treatment plan of the target special situation problem;
[0025] When the average similarity between the solution for the target special situation problem and the solution included in each of the associated special situation knowledge data does not meet a preset similarity threshold, a second target number of associated special situation knowledge data associated with the target special situation problem is retrieved from the special situation knowledge vector library, wherein the second target number is equal to the first target number plus one.
[0026] In some embodiments of the present disclosure, the method further includes:
[0027] Acquire flight safety standard knowledge document data and historical flight accident data, wherein the historical flight accident data includes accident information and disposal plans;
[0028] Determining the associated documents of the historical flight accident data according to the disposal plan of the historical flight accident data and the associated relationship of the flight safety standard knowledge document data;
[0029] A special situation knowledge vector library is formed after the accident information, handling plans and related documents of the historical flight accident data are vectorized.
[0030] In a second aspect, according to the content of the present disclosure, a flight safety special situation handling device is provided, comprising:
[0031] A related special situation knowledge data determination module is used to retrieve a first target number of related special situation knowledge data associated with the target special situation problem from a special situation knowledge vector library according to the target special situation problem, wherein the special situation knowledge vector library includes a plurality of special situation knowledge data, and the special situation knowledge data includes accident information, disposal solutions and related documents;
[0032] A prompt word generation module, used for generating prompt words for the target special situation problem according to the target special situation problem, the first target number of associated special situation knowledge data and the prompt word template;
[0033] A treatment plan generating module, used for generating a treatment plan for the target special situation problem according to the prompt word of the target special situation problem;
[0034] The target handling solution determination module is used to determine the target handling solution of the target special situation problem based on the average similarity between the handling solution of the target special situation problem and the handling solutions included in each of the associated special situation knowledge data.
[0035] In a third aspect, according to the present disclosure, there is provided a computer device, including:
[0036] one or more processors;
[0037] a storage device for storing one or more programs,
[0038] When the one or more programs are executed by the one or more processors, the one or more processors implement any method as described in the first aspect.
[0039] In a fourth aspect, according to the content of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored, and when the program is executed by a processor, it implements any method as described in the first aspect.
[0040] The flight safety special situation handling method, device, equipment and medium provided by the disclosed embodiment first retrieve the first target number of associated special situation knowledge data associated with the target special situation problem from the special situation knowledge vector library according to the target special situation problem; then generate the prompt word of the target special situation problem according to the target special situation problem, the first target number of associated special situation knowledge data and the prompt word template; and generate the handling plan of the target special situation problem according to the prompt word of the target special situation problem; finally, determine the target handling plan of the target special situation problem according to the average similarity between the handling plan of the target special situation problem and the associated special situation knowledge data. Since the prompt word is generated by the target special situation problem and the associated special situation knowledge data that has an associated relationship with the target special situation problem, the handling plan of the target special situation problem generated based on the prompt word has a high accuracy, and the handling plan generated by the large language model is further verified to further ensure the accuracy of the result.
[0041] The above description is only an overview of the technical solution of the embodiment of the present application. In order to more clearly understand the technical means of the embodiment of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the embodiment of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings of the embodiments will be briefly described below. It should be noted that the drawings described below only relate to some embodiments of the present disclosure, but are not intended to limit the present disclosure, wherein:
[0043] Figure 1 It is a flowchart of a method for handling a flight safety special situation provided by an embodiment of the present disclosure;
[0044] Figure 2 It is a structural schematic diagram of a flight safety special situation handling device provided by an embodiment of the present disclosure;
[0045] Figure 3 It is a structural diagram of a computer device provided in an embodiment of the present disclosure.
[0046] In the drawings, reference numerals with the same last two digits correspond to the same elements. It should be noted that the elements in the drawings are schematic and not drawn to scale. DETAILED DESCRIPTION
[0047] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution of the embodiments of the present disclosure will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present disclosure, rather than all of the embodiments. Based on the described embodiments of the present disclosure, all other embodiments obtained by those skilled in the art without creative work also fall within the scope of protection of the present disclosure.
[0048] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by a person skilled in the art to which the subject matter of the present disclosure belongs. It will be further understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the specification and the relevant art, and will not be interpreted in an idealized or overly formal form unless otherwise explicitly defined herein. As used herein, a statement that two or more parts are "connected" or "coupled" together shall mean that the parts are joined together directly or through one or more intermediate components.
[0049] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase "embodiments" in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.
[0050] The term "and / or" in this article is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists, A and B exist at the same time, and B exists. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.
[0051] Furthermore, in all embodiments of the present disclosure, terms such as “first” and “second” are used only to distinguish one component (or a part of a component) from another component (or another part of a component).
[0052] In the description of the present application, unless otherwise specified, "plurality" means more than two (including two), and similarly, "multiple groups" means more than two groups (including two).
[0053] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.
[0054] Based on the problems existing in the prior art, the present disclosure provides a method for handling special flight safety situations. Figure 1 is a flowchart of a method for handling special flight safety situations provided by an embodiment of the present disclosure, such as Figure 1 As shown in the figure, the flight safety special situation handling methods include:
[0055] S110. According to the target special situation problem, a first target number of associated special situation knowledge data associated with the target special situation problem is retrieved from the special situation knowledge vector library.
[0056] Among them, the special situation knowledge vector library includes multiple special situation knowledge data, and the special situation knowledge data includes accident information, handling plans and related documents.
[0057] In a specific implementation, in the flight safety special situation handling method provided by the embodiment of the present disclosure, when a target special situation problem is received, firstly, based on the target special situation problem, a first target number of associated special situation knowledge data associated with the target special situation problem is retrieved from the special situation knowledge vector library.
[0058] In a specific implementation manner, before executing step S110, it includes: constructing a special situation knowledge vector library, wherein the specific implementation process of constructing the special situation knowledge vector library includes: obtaining flight safety standard knowledge document data and historical flight accident data, wherein the historical flight accident data includes accident information and disposal plans; determining the associated documents of the historical flight accident data according to the association between the disposal plans of the historical flight accident data and the flight safety standard knowledge document data; forming special situation knowledge data after vectorizing the accident information, disposal plans and associated documents of the historical flight accident data, and the special situation knowledge data corresponding to different historical flight accident data constitute the special situation knowledge vector library.
[0059] The specific process of vectorizing the accident information, disposal plan and related documents of historical flight accident data is as follows: select the corresponding vectorization method such as Bag-of-Words, Word2Vec, GloVe, Context Embedding Model (BERT, GPT), etc. to vectorize the document, and finally form the vectorized representation of the document, that is, text vector data. Then, based on the text vector data, use the vector database (Faiss, Milvus, Pinecone, etc.) for storage to form a special situation knowledge vector library. At the same time, in order to further improve the retrieval efficiency, the special situation knowledge vector library is indexed, that is, the accident information, disposal plan and related documents corresponding to a historical flight accident data constitute a special situation knowledge data, and the special situation knowledge data corresponds to a vector library index value.
[0060] Specifically, the accident information is used to describe the accident case, the disposal plan is used to describe the disposal plan corresponding to the accident, and the associated document is used to describe the associated document information and path corresponding to the accident.
[0061] Specifically, according to the target special situation problem, a first target number of related special situation knowledge data associated with the target special situation problem is retrieved from the special situation knowledge vector library, including: vectorizing the target special situation problem to obtain a target special situation problem vector; performing similarity matching on the target special situation problem vector and the accident information included in each special situation knowledge data to obtain a similarity value; after sorting the similarity values of the target special situation problem vector and the accident information included in each special situation knowledge data in reverse order, selecting the first target number of special situation knowledge data to form the related special situation knowledge data.
[0062] Since the special situation knowledge data in the special situation knowledge vector library are stored in the form of text vector data, and the accident information in the special situation knowledge data describes the accident case, therefore, in the process of determining the associated special situation knowledge data, the target special situation problem is first vectorized to obtain the target special situation problem vector, and then the target special situation problem vector is matched with each special situation knowledge data in the special situation knowledge vector library for similarity, that is, the target special situation problem vector is matched with the accident information corresponding to each special situation knowledge data for similarity, and according to the similarity value of the target special situation problem vector and the accident information corresponding to each special situation knowledge data, the first target number of special situation knowledge data is selected to form the associated special situation knowledge data.
[0063] S120, generating prompt words for the target special situation problem according to the target special situation problem, the first target number of related special situation knowledge data and the prompt word template.
[0064] In a specific implementation, prompt words for the target special situation problem are generated based on the target special situation problem, the associated special situation knowledge data of the first target quantity, and a prompt word template, including: determining the special situation handling context input of the target special situation problem based on the associated documents included in the associated special situation knowledge data of the first target quantity; filling the target special situation problem and the special situation handling context input into the prompt word template to generate prompt words for the target special situation problem.
[0065] Specifically, after the associated special situation knowledge data is acquired in step S110, the special situation handling context input of the target special situation problem is determined based on the associated documents corresponding to the associated special situation knowledge data, and then the target special situation problem and the special situation handling context input are filled into the prompt word template to generate the prompt word for the target special situation problem.
[0066] S130. Generate a solution to the target special situation problem based on the prompt word of the target special situation problem.
[0067] The prompt words of the target special situation problem generated in step S120 are input into the large language model, and a solution to the target special situation problem is generated based on the large language model.
[0068] Specifically, the large language model is mainly a model generated by training special situation handling question and answer pair data. By training the parameters of the large language model with the special situation handling question and answer pair data, a large language model with an accuracy that meets the preset accuracy is obtained.
[0069] The specific process of generating a large language model based on special situation handling question and answer data includes: converting the data format of the special situation handling question and answer data to obtain special situation handling question and answer input data suitable for the large language model, and setting the corresponding learning rate of the large language model, the number of samples processed during each training, the number of samples processed during each evaluation, the weight decay rate, the maximum number of checkpoints saved, the number of training rounds, the step interval of saving the model, the step interval of the evaluation model, the evaluation index of selecting the best model, etc., and further loading the training data set, the evaluation data set, the model parameters, etc. to train the model. When the large language model meets the convergence conditions, the large language model generated by the training is used as the large language model of this application.
[0070] Among them, the special situation handling question and answer data are mainly compiled by commanders / pilots and other relevant professionals. The special situation handling question and answer data include special situation questions or instructions, special situation handling context input, special situation handling plan output and problem sources. Special situation questions or instructions are questions asked or instructions issued by commanders or pilots to the big model based on special situations occurring in flight safety. Special situation handling context input is an optional input provided as the context of special situation handling questions or instructions. Special situation handling plan output is a handling plan for the expected output of special situation questions or instructions and special situation handling context input. The problem source identifies that the question and answer pair is based on the commander / pilot's knowledge and experience or relevant historical flight accident data.
[0071] Since the prompt words of the generated target special situation problem include the associated documents of the associated special situation knowledge data and the target special situation problem, and the associated documents of the associated special situation knowledge data include the associated document information and paths, the accuracy of the handling solution of the target special situation problem generated based on the prompt words of the target special situation problem is higher than that of the handling solution generated directly based on the target special situation problem.
[0072] S140. Determine a target handling solution for the target special situation problem based on the average similarity between the handling solution for the target special situation problem and the associated special situation knowledge data.
[0073] In a specific implementation, a target handling solution for a target special situation problem is determined based on the average similarity between the handling solution for a target special situation problem and the associated special situation knowledge data, including: sequentially determining the similarity between the handling solution for the target special situation problem and the handling solutions included in each associated special situation knowledge data; summing up the similarities between the handling solution for the target special situation problem and the handling solutions corresponding to each associated special situation knowledge data and taking the average value to obtain the average similarity between the handling solution for the target special situation problem and the handling solutions corresponding to each associated special situation knowledge data; determining the target handling solution for the target special situation problem based on the relationship between the average similarity between the handling solution for the target special situation problem and the handling solutions corresponding to each associated special situation knowledge data and a preset similarity threshold.
[0074] Specifically, since the associated special situation knowledge data determined in step S110 includes the first target quantity, after the large language model outputs the solution to the target special situation problem in step S130, the solution to the target special situation problem output by the large language model is matched with the solution included in each associated special situation knowledge data in turn to obtain the similarity, and then the similarities obtained between the solution included in each associated special situation knowledge data and the solution to the target special situation problem output by the large language model are summed and averaged to obtain the average similarity between the solution to the target special situation problem output by the large language model and the solution corresponding to each associated special situation knowledge data, and finally, by evaluating the relationship between the average similarity and the preset similarity, it is determined whether the solution output by the large language model is the solution to the target special situation problem.
[0075] Specifically, if the average similarity between the solution of the target special situation problem and the solution corresponding to each related special situation knowledge data meets the preset similarity threshold, the solution generated by the large language model based on the prompt word of the target special situation problem is the target solution of the target special situation problem; if the average similarity between the solution of the target special situation problem and the solution corresponding to each related special situation knowledge data does not meet the preset similarity threshold, a second target number of related special situation knowledge data associated with the target special situation problem is retrieved from the special situation knowledge vector library, where the second target number is equal to the first target number plus one.
[0076] That is, the target handling solution for the target special situation problem is determined by comparing the relationship between the average similarity between the handling solution for the target special situation problem and the handling solutions corresponding to each related special situation knowledge data and the similarity threshold. If the average similarity between the handling solution for the target special situation problem and the handling solutions corresponding to each related special situation knowledge data meets the preset similarity threshold, the handling solution generated by the large language model according to the prompt word for the target special situation problem is the target handling solution for the target special situation problem. If the average similarity between the handling solution for the target special situation problem and the handling solutions corresponding to each related special situation knowledge data does not meet the preset similarity threshold, the accuracy of the handling solution for the target special situation problem generated based on the prompt word is ensured by increasing the number of related special situation knowledge data retrieved from the special situation knowledge vector library and increasing the description of the generated prompt words.
[0077] The flight safety special situation handling method provided by the embodiment of the present disclosure first retrieves the first target number of associated special situation knowledge data associated with the target special situation problem from the special situation knowledge vector library according to the target special situation problem; then generates the prompt word of the target special situation problem according to the target special situation problem, the first target number of associated special situation knowledge data and the prompt word template; and generates the handling plan of the target special situation problem according to the prompt word of the target special situation problem; finally, determines the target handling plan of the target special situation problem according to the average similarity between the handling plan of the target special situation problem and the associated special situation knowledge data. Since the prompt word is generated by the target special situation problem and the associated special situation knowledge data that has an associated relationship with the target special situation problem, the handling plan of the target special situation problem generated based on the prompt word has a high accuracy, and the handling plan generated by the large language model is further verified to further ensure the accuracy of the result.
[0078] Based on the above embodiments, the present disclosure also provides a flight safety special situation handling device. Figure 2 is a schematic diagram of the structure of the flight safety special situation handling device provided by the embodiment of the present disclosure, such as Figure 2 As shown in the figure, the flight safety special situation handling device includes:
[0079] The associated special situation knowledge data determining module 210 is used to retrieve a first target number of associated special situation knowledge data associated with the target special situation problem from the special situation knowledge vector library according to the target special situation problem, wherein the special situation knowledge vector library includes a plurality of special situation knowledge data, and the special situation knowledge data includes accident information, disposal solutions and associated documents;
[0080] A prompt word generation module 220, for generating prompt words for the target special situation problem according to the target special situation problem, the first target number of associated special situation knowledge data and the prompt word template;
[0081] A treatment plan generating module 230 is used to generate a treatment plan for a target special situation problem according to the prompt words of the target special situation problem;
[0082] The target handling solution determination module 240 is used to determine the target handling solution for the target special situation problem according to the average similarity between the handling solution for the target special situation problem and the handling solutions included in each associated special situation knowledge data.
[0083] The flight safety special situation handling device provided by the disclosed embodiment first retrieves the first target number of associated special situation knowledge data associated with the target special situation problem from the special situation knowledge vector library according to the target special situation problem; then generates the prompt word of the target special situation problem according to the target special situation problem, the first target number of associated special situation knowledge data and the prompt word template; and generates the handling plan of the target special situation problem according to the prompt word of the target special situation problem; finally, determines the target handling plan of the target special situation problem according to the average similarity between the handling plan of the target special situation problem and the associated special situation knowledge data. Since the prompt word is generated by the target special situation problem and the associated special situation knowledge data that has an associated relationship with the target special situation problem, the handling plan of the target special situation problem generated based on the prompt word has a high accuracy, and the handling plan generated by the large language model is further verified to further ensure the accuracy of the result.
[0084] In a specific implementation, the step of retrieving a first target number of associated special situation knowledge data associated with the target special situation problem from a special situation knowledge vector library according to the target special situation problem includes:
[0085] Performing vectorization processing on the target special situation problem to obtain a target special situation problem vector;
[0086] Respectively performing similarity matching between the target special situation problem vector and the accident information included in each of the special situation knowledge data to obtain a similarity value;
[0087] After the similarity values of the target special situation problem vector and the accident information included in each of the special situation knowledge data are sorted in reverse order, the special situation knowledge data of the first target quantity are selected to form the associated special situation knowledge data.
[0088] In a specific implementation, generating a prompt word for the target special situation problem according to the target special situation problem, the first target number of associated special situation knowledge data and a prompt word template includes:
[0089] Determining a special situation handling context input for the target special situation problem according to the associated documents included in the first target number of associated special situation knowledge data;
[0090] The target special situation problem and the special situation handling context are input into a prompt word template to generate a prompt word for the target special situation problem.
[0091] In a specific implementation, generating a solution to the target special situation problem according to the prompt word of the target special situation problem includes:
[0092] The prompt words of the target special situation problem are input into the large language model, and the prompt words of the target special situation problem are analyzed and identified based on the large language model to obtain a solution to the target special situation problem.
[0093] In a specific implementation, determining a target solution for the target special situation problem according to the average similarity between the solution for the target special situation problem and the solution included in each of the associated special situation knowledge data includes:
[0094] Determining in sequence the similarity between the solution for the target special situation problem and the solution included in each of the associated special situation knowledge data;
[0095] The similarities between the treatment plan of the target special situation problem and the treatment plan corresponding to each of the associated special situation knowledge data are summed and averaged to obtain the average similarity between the treatment plan of the target special situation problem and the treatment plan included in each of the associated special situation knowledge data;
[0096] The target handling solution for the target special situation problem is determined based on the relationship between the average similarity between the handling solution for the target special situation problem and the handling solutions included in each of the associated special situation knowledge data and a preset similarity threshold.
[0097] In a specific implementation, determining a target solution for the target special situation problem according to the average similarity between the solution for the target special situation problem and the solution included in each of the associated special situation knowledge data includes:
[0098] When the average similarity between the treatment plan of the target special situation problem and the treatment plans included in each of the associated special situation knowledge data meets the preset similarity threshold, the treatment plan generated according to the prompt word of the target special situation problem is the target treatment plan of the target special situation problem;
[0099] When the average similarity between the solution for the target special situation problem and the solution included in each of the associated special situation knowledge data does not meet a preset similarity threshold, a second target number of associated special situation knowledge data associated with the target special situation problem is retrieved from the special situation knowledge vector library, wherein the second target number is equal to the first target number plus one.
[0100] In a specific embodiment, the method further comprises:
[0101] Acquire flight safety standard knowledge document data and historical flight accident data, wherein the historical flight accident data includes accident information and disposal plans;
[0102] Determining the associated documents of the historical flight accident data according to the disposal plan of the historical flight accident data and the associated relationship of the flight safety standard knowledge document data;
[0103] A special situation knowledge vector library is formed after the accident information, handling plans and related documents of the historical flight accident data are vectorized.
[0104] The present application also provides a computer device. Figure 3 , Figure 3 This is a basic structural block diagram of the computer device in this embodiment.
[0105] The computer device includes a memory 510 and a processor 520 that are connected to each other through a system bus. It should be noted that the figure only shows a computer device with components 510-520, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.
[0106] Computer devices can be computing devices such as desktop computers, notebooks, PDAs, and cloud servers. Computer devices can interact with users through keyboards, mice, remote controls, touch pads, or voice control devices.
[0107] The memory 510 includes at least one type of readable storage medium, and the readable storage medium includes a non-volatile memory or a volatile memory, such as a flash memory, a hard disk, a multimedia card, a card-type memory (e.g., SD or DX memory, etc.), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disk, etc., and the RAM may include a static RAM or a dynamic RAM. In some embodiments, the memory 510 may be an internal storage unit of a computer device, such as a hard disk or a memory of the computer device. In other embodiments, the memory 510 may also be an external storage device of a computer device, such as a plug-in hard disk, a smart memory card (SMC), a secure digital (SD) card, or a flash card (Flash Card), etc., equipped on the computer device. Of course, the memory 510 may also include both an internal storage unit of the computer device and an external storage device thereof. In this embodiment, the memory 510 is generally used to store an operating system and various application software installed on the computer device, such as the program code of the above method, etc. In addition, the memory 510 may also be used to temporarily store various data that have been output or are to be output.
[0108] The processor 520 is generally used to perform the overall operation of the computer device. In this embodiment, the memory 510 is used to store program codes or instructions, the program code includes computer operation instructions, and the processor 520 is used to execute the program codes or instructions stored in the memory 510 or process data, such as running the program code of the above method.
[0109] In this article, the bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus system 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 it does not mean that there is only one bus or one type of bus.
[0110] Another embodiment of the present application also provides a computer-readable medium, which may be a computer-readable signal medium or a computer-readable medium. A processor in a computer reads a computer-readable program code stored in the computer-readable medium, so that the processor can execute the functional actions specified in each step or a combination of steps in the above method; and generate a device for implementing the functional actions specified in each block or a combination of blocks in the block diagram.
[0111] Computer-readable media include but are not limited to electronic, magnetic, optical, electromagnetic, infrared memory or semiconductor systems, devices or apparatuses, or any appropriate combination of the foregoing, the memory is used to store program codes or instructions, the program codes include computer operating instructions, and the processor is used to execute the program codes or instructions of the above methods stored in the memory.
[0112] For the definitions of memory and processor, please refer to the description of the aforementioned computer device embodiment and will not be repeated here.
[0113] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of modules or units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.
[0114] Each functional unit or module in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.
[0115] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) or a processor (processor) to perform all or part of the steps of each embodiment method of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), disk or optical disk and other media that can store program code.
[0116] Unless the context clearly indicates otherwise, the singular form of the words used herein and in the appended claims includes the plural and vice versa. Thus, when referring to the singular, the plural form of the corresponding term is generally included. Similarly, the words "comprise" and "include" are to be interpreted as inclusive rather than exclusive. Likewise, the terms "include" and "or" should be interpreted as inclusive unless such interpretation is expressly prohibited herein. Where the term "example" is used herein, particularly when it is located after a group of terms, the "example" is merely exemplary and illustrative and should not be considered exclusive or comprehensive.
[0117] Further aspects and scopes of adaptability become apparent from the description provided herein. It should be understood that various aspects of the present application can be implemented individually or in combination with one or more other aspects. It should also be understood that the description and specific embodiments herein are intended for purposes of illustration only and are not intended to limit the scope of the present application.
[0118] Several embodiments of the present disclosure are described in detail above, but it is obvious that those skilled in the art can make various modifications and variations to the embodiments of the present disclosure without departing from the spirit and scope of the present disclosure. The protection scope of the present disclosure is defined by the attached claims.
Claims
1. A method for handling special flight safety situations, characterized in that: include: According to the target special situation problem, a first target number of associated special situation knowledge data associated with the target special situation problem is retrieved from a special situation knowledge vector library, wherein the special situation knowledge vector library includes a plurality of special situation knowledge data, and the special situation knowledge data includes accident information, disposal plans and associated documents; Generate a prompt word for the target special situation question according to the target special situation question, the first target number of associated special situation knowledge data and a prompt word template; Generate a solution to the target special situation problem according to the prompt word of the target special situation problem; A target handling solution for the target special situation problem is determined based on an average similarity between the handling solution for the target special situation problem and the handling solutions included in each of the associated special situation knowledge data.
2. The method according to claim 1, characterized in that The step of retrieving a first target number of associated special situation knowledge data associated with the target special situation problem from a special situation knowledge vector library according to the target special situation problem comprises: Performing vectorization processing on the target special situation problem to obtain a target special situation problem vector; Respectively performing similarity matching between the target special situation problem vector and the accident information included in each of the special situation knowledge data to obtain a similarity value; After the similarity values of the target special situation problem vector and the accident information included in each of the special situation knowledge data are sorted in reverse order, the special situation knowledge data of the first target quantity are selected to form the associated special situation knowledge data.
3. The method according to claim 1, characterized in that The step of generating a prompt word for the target special situation problem according to the target special situation problem, the first target number of associated special situation knowledge data and a prompt word template comprises: Determining a special situation handling context input for the target special situation problem according to the associated documents included in the first target number of associated special situation knowledge data; The target special situation problem and the special situation handling context are input into a prompt word template to generate a prompt word for the target special situation problem.
4. The method according to claim 1, characterized in that Generating a solution to the target special situation problem according to the prompt word of the target special situation problem includes: The prompt words of the target special situation problem are input into the large language model, and the prompt words of the target special situation problem are analyzed and identified based on the large language model to obtain a solution to the target special situation problem.
5. The method according to claim 1, characterized in that Determining a target solution for the target special situation problem according to the average similarity between the solution for the target special situation problem and the solution included in each of the associated special situation knowledge data comprises: Determining in sequence the similarity between the solution for the target special situation problem and the solution included in each of the associated special situation knowledge data; The similarities between the treatment plan of the target special situation problem and the treatment plan corresponding to each of the associated special situation knowledge data are summed and averaged to obtain the average similarity between the treatment plan of the target special situation problem and the treatment plan included in each of the associated special situation knowledge data; The target handling solution for the target special situation problem is determined based on the relationship between the average similarity between the handling solution for the target special situation problem and the handling solutions included in each of the associated special situation knowledge data and a preset similarity threshold.
6. The method according to claim 1 or 5, characterized in that: Determining a target solution for the target special situation problem according to the average similarity between the solution for the target special situation problem and the solution included in each of the associated special situation knowledge data comprises: When the average similarity between the treatment plan of the target special situation problem and the treatment plans included in each of the associated special situation knowledge data meets the preset similarity threshold, the treatment plan generated according to the prompt word of the target special situation problem is the target treatment plan of the target special situation problem; When the average similarity between the solution for the target special situation problem and the solution included in each of the associated special situation knowledge data does not meet a preset similarity threshold, a second target number of associated special situation knowledge data associated with the target special situation problem is retrieved from the special situation knowledge vector library, wherein the second target number is equal to the first target number plus one.
7. The method according to claim 1, characterized in that The method further comprises: Acquire flight safety standard knowledge document data and historical flight accident data, wherein the historical flight accident data includes accident information and disposal plans; Determining the associated documents of the historical flight accident data according to the disposal plan of the historical flight accident data and the associated relationship of the flight safety standard knowledge document data; A special situation knowledge vector library is formed after the accident information, handling plans and related documents of the historical flight accident data are vectorized.
8. A flight safety special situation handling device, characterized in that: include: A related special situation knowledge data determination module is used to retrieve a first target number of related special situation knowledge data associated with the target special situation problem from a special situation knowledge vector library according to the target special situation problem, wherein the special situation knowledge vector library includes a plurality of special situation knowledge data, and the special situation knowledge data includes accident information, disposal solutions and related documents; A prompt word generation module, used for generating prompt words for the target special situation problem according to the target special situation problem, the first target number of associated special situation knowledge data and the prompt word template; A treatment plan generating module, used for generating a treatment plan for the target special situation problem according to the prompt word of the target special situation problem; The target handling solution determination module is used to determine the target handling solution of the target special situation problem based on the average similarity between the handling solution of the target special situation problem and the handling solutions included in each of the associated special situation knowledge data.
9. A computer device, characterized in that: include: one or more processors; a 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 according to any one of claims 1 to 7.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.
Citation Information
Patent Citations
Aviation safety report analysis and evaluation method based on text similarity model
CN115186660A
Knowledge retrieval method, equipment, storage medium and device
CN117763126A
Knowledge base response method and system based on large language model
CN117951249A
Knowledge retrieval enhancement generation method and system based on large language model
CN118394890A
