Insurance policy risk identification method and device, computer equipment and storage medium

Abstract

The invention belongs to the field of artificial intelligence and financial science and technology, and relates to an insurance policy risk identification method and device, computer equipment and a storage medium, and the method comprises the steps: obtaining a to-be-identified special text and an insurance clause text; extracting a special responsibility set from the special text, and extracting a responsibility clause set and a free clause set from the insurance clause text; calculating a difference set of the special responsibility set and the responsibility clause set and an intersection of the special responsibility set and the free clause set by adopting a semantic understanding model; determining an extended responsibility range according to the difference set, and determining an exempt responsibility range according to the intersection; and according to the responsibility types of the expanded responsibility range and the exempted responsibility range, determining the risk level of the special responsibility terms contained in each responsibility type, and determining the prompt content of the special responsibility terms. In addition, the invention also relates to a block chain technology, and data such as a special text and an insurance clause text can be stored in a block chain. According to the invention, efficient risk identification can be carried out for the special clauses.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence and financial technology, and in particular to a method, device, computer equipment and storage medium for identifying insurance policy risks. Background Art

[0002] In the process of carrying out insurance business, it is often necessary to use special clauses (i.e. special agreed liability clauses) to meet additional agreements with customers. These special clauses include fixed special clauses, semi-fixed special clauses and free-edited special clauses, which have legal effect beyond the stock clauses. Among them, free-edited special clauses, due to their high degree of freedom and strong legal effect, often contain protection content inconsistent with the stock clauses. Although this liability extension is flexible, it also brings the risk of claims leakage. At present, the risk review of free-edited special clauses mainly relies on the evaluation of each case by professional underwriters. This process is time-consuming, labor-intensive and inefficient. Summary of the invention

[0003] The purpose of the embodiments of the present application is to propose a policy risk identification method, device, computer equipment and storage medium, which can efficiently identify risks for special terms.

[0004] In order to solve the above technical problems, the embodiment of the present application provides a method for identifying insurance policy risks, which adopts the following technical solutions:

[0005] Obtaining the special contract text to be identified and the insurance clause text corresponding to the special contract text;

[0006] Using a preset liability set extraction rule, extract a special liability set from the special contract text, and extract a liability clause set and an exemption clause set from the insurance clause text, wherein the special liability set, the liability clause set and the exemption clause set are all composed of minimum semantic units;

[0007] Input the special liability set, liability clause set and exemption clause set into the pre-trained semantic understanding model, and use the dot product attention mechanism to calculate the difference between the special liability set and the liability clause set, and the intersection between the special liability set and the exemption clause set;

[0008] Determine the extended liability scope corresponding to the special contract text according to the difference set, and determine the exemption liability scope corresponding to the special contract text according to the intersection set;

[0009] According to the liability types of the extended liability scope and the exempted liability scope, the risk level of the special liability clauses included in each liability type is determined, and the prompt content of the special liability clauses is determined according to the risk level.

[0010] In order to solve the above technical problems, the embodiment of the present application further provides a device for identifying insurance policy risks, which adopts the following technical solution:

[0011] An acquisition module, used to acquire the special contract text to be identified and the insurance clause text corresponding to the special contract text;

[0012] An extraction module, used to extract a special liability set from the special contract text, and extract a liability clause set and a disclaimer clause set from the insurance clause text by using a preset liability set extraction rule, wherein the special liability set, the liability clause set and the disclaimer clause set are all composed of minimum semantic units;

[0013] A calculation module, used for inputting the special liability set, the liability clause set and the exemption clause set into a pre-trained semantic understanding model, and using a dot product attention mechanism to calculate the difference between the special liability set and the liability clause set, and the intersection between the special liability set and the exemption clause set;

[0014] A determination module, used to determine the extended liability scope corresponding to the special contract text according to the difference set, and determine the exemption liability scope corresponding to the special contract text according to the intersection set;

[0015] A grading module is used to determine the risk level of the special liability clauses included in each liability type according to the liability types of the extended liability scope and the exempted liability scope, and to determine the prompt content of the special liability clauses according to the risk level.

[0016] In order to solve the above-mentioned technical problems, an embodiment of the present application also provides a computer device, which adopts the technical solution described below: the computer device includes a memory and a processor, the memory stores computer-readable instructions, and the processor implements the steps of the insurance policy risk identification method as described in any one of the above items when executing the computer-readable instructions.

[0017] In order to solve the above-mentioned technical problems, an embodiment of the present application also provides a computer-readable storage medium, which adopts the technical solution described below: the computer-readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the steps of the insurance policy risk identification method as described in any one of the above items are implemented.

[0018] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0019] The embodiment of the present application adopts preset liability set extraction rules to accurately extract liability sets from special texts and insurance clauses, including special liability sets, liability clause sets and exemption clause sets; by adopting a pre-trained semantic understanding model, it can deeply understand the semantic information in the special texts and insurance clauses, and adopt a dot product attention mechanism to calculate the difference between the special liability set and the liability clause set, as well as the intersection of the special liability set and the exemption clause set, so as to accurately identify the extended liability scope and the exemption liability scope corresponding to the special text. By abstracting the semantic problem into a mathematical problem and processing the liability definition in a set-based manner, compared with traditional semantic comparison reasoning, not only is the working mechanism clear, easy to trace back and explain, but also the accuracy and efficiency of recognition are improved; by determining the risk level of the special liability clauses contained in each liability type according to the liability type of the extended liability scope and the exemption liability scope, and determining the prompt content of the special liability clause according to the risk level, the business process is made more scientific and reasonable, and the hierarchical management and control of risks can be achieved, thereby improving the efficiency and effectiveness of business processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0021] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0022] Figure 2 It is a flowchart of an embodiment of the insurance policy risk identification method of the present application;

[0023] Figure 3 It is a structural schematic diagram of an embodiment of the insurance policy risk identification device of the present application;

[0024] Figure 4 It is a structural diagram of an embodiment of a computer device of the present application. DETAILED DESCRIPTION

[0025] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0026] 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 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.

[0027] 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.

[0028] like Figure 1 As shown, the system architecture 100 may include a terminal device 101, a network 102 and a server 103. The terminal device 101 may be a laptop 1011, a tablet computer 1012 or a mobile phone 1013. The network 102 is used to provide a medium for a communication link between the terminal device 101 and the server 103. The network 102 may include various connection types, such as wired, wireless communication links or optical fiber cables.

[0029] The user can use the terminal device 101 to interact with the server 103 through the network 102 to receive or send messages, etc. Various communication client applications can be installed on the terminal device 101, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0030] The terminal device 101 can be any electronic device with a display screen and supporting web browsing. In addition to a laptop computer 1011, a tablet computer 1012 or a mobile phone 1013, the terminal device 101 can also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 (Moving Picture Experts Group Audio Layer IV), a laptop computer, a desktop computer, etc.

[0031] The server 103 may be a server that provides various services, such as a background server that provides support for a web page displayed on the terminal device 101 .

[0032] It should be noted that the insurance policy risk identification method provided in the embodiment of the present application is generally executed by a server / terminal device, and accordingly, the insurance policy risk identification device is generally set in the server / terminal device.

[0033] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to the implementation requirements.

[0034] Continue to refer Figure 2 , showing a flow chart of an embodiment of the question-answering method according to the present application. The insurance policy risk identification method comprises the following steps:

[0035] Step S201, obtaining a special contract text to be identified and an insurance clause text corresponding to the special contract text;

[0036] In this embodiment, the insurance policy risk identification method is executed on an electronic device (eg Figure 1 The server / terminal device shown in the figure) can obtain the special contract text to be identified and the insurance clause text corresponding to the special contract text through a wired connection method or a wireless connection method. It should be noted that the above wireless connection method may include but is not limited to 2G / 3G / 4G / 5G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other wireless connection methods currently known or to be developed in the future.

[0037] Specifically, obtain the special text to be identified and the insurance clause text corresponding to the special text. Special text refers to the special liability clauses agreed upon by the two parties (insurance company and policyholder) outside the main contract based on specific needs or circumstances during the insurance contract signing process. The special liability clauses may involve specific coverage, liability limitations, exemption conditions, claims conditions, etc. The insurance clause text is the core content of the insurance contract, which stipulates the rights and obligations between the insurance company and the policyholder, including insurance liability, exemption clauses, claims process, etc. The insurance clause text may include the main insurance clause text and the supplementary insurance clause text, among which the main insurance clause text is the clause part that stipulates the basic content of the insurance contract and the basic rights and obligations of both parties; the supplementary insurance clause text is the clause part that supplements or expands the main insurance clause text.

[0038] Step S202, using a preset liability set extraction rule, extracting a special liability set from the special contract text, and extracting a liability clause set and an exemption clause set from the insurance clause text, wherein the special liability set, liability clause set and exemption clause set are all composed of minimum semantic units;

[0039] In this embodiment, a liability set extraction rule is preset, and the liability set extraction rule is formulated according to the characteristics of the insurance clauses, and may include a special liability extraction rule, a liability clause extraction rule, an exemption clause extraction rule, etc. The preset liability set extraction rule is obtained, and the special text is analyzed sentence by sentence using the liability set extraction rule, split into the smallest semantic unit, and the special liability set is extracted; the insurance clause text is analyzed sentence by sentence using the liability set extraction rule, split into the smallest semantic unit, and the liability clause set and the exemption clause set are extracted.

[0040] In some optional implementations, the above-mentioned using a preset liability set extraction rule to extract a special liability set from the special contract text, and extracting a liability clause set and a disclaimer clause set from the insurance clause text may include the following steps:

[0041] Obtaining a preset responsibility description template, wherein the responsibility description template includes a responsibility tag related to the responsibility content, wherein the responsibility tag is set according to a preset responsibility set extraction rule, and wherein the responsibility tag includes a special responsibility tag, a responsibility scope tag, and an exemption scope tag;

[0042] Specifically, a preset liability description template is obtained. The liability description template may include a liability tag related to the liability content. The liability tag is set according to a preset liability set extraction rule. The liability set extraction rule may include a special liability extraction rule, a liability clause extraction rule, and an exemption clause extraction rule. The liability tag may include a special liability tag, a liability scope tag, and an exemption scope tag. Each liability tag corresponds to a specific liability content, such as the liability object (insured, policyholder), liability event (accidental injury, disease diagnosis), compensation amount, compensation ratio, compensation conditions, and the like.

[0043] Using the responsibility description template, extracting the minimum semantic unit matching the special responsibility label from the special text as a special responsibility set;

[0044] Specifically, the acquired responsibility description template is applied to the special text, the special text is split into multiple minimum semantic units, and the special responsibility label in the responsibility description template is compared one by one with each of the split minimum semantic units, and the minimum semantic unit matching the special responsibility label is extracted as the special responsibility set.

[0045] The liability description template is used to extract the smallest semantic unit that matches the liability scope label from the insurance clause text as a liability clause set, and the smallest semantic unit that matches the exemption scope label is extracted from the insurance clause text as an exemption clause set.

[0046] Specifically, the acquired liability description template is applied to the insurance clause text, the insurance clause text is split into multiple minimum semantic units, and the liability scope label in the liability description template is compared with each of the split minimum semantic units one by one, and the minimum semantic units matching the liability scope label are extracted as the liability clause set. The exemption scope label in the liability description template is compared with each of the split minimum semantic units one by one, and the minimum semantic units matching the exemption scope label are extracted as the exemption clause set.

[0047] The embodiment of the present application sets a liability description template according to liability set extraction rules, compares the liability labels and the various text contents in the liability description template, can accurately identify the smallest semantic unit that matches the liability label, and efficiently extract key liability information from the special text and insurance clause text, thereby avoiding semantic misunderstandings and omissions.

[0048] Step S203, inputting the special liability set, liability clause set and exemption clause set into a pre-trained semantic understanding model, and using a dot product attention mechanism to calculate the difference between the special liability set and the liability clause set, and the intersection between the special liability set and the exemption clause set;

[0049] In this embodiment, the semantic understanding model is a machine learning model for understanding and processing text semantic information. The model is built based on deep learning technology and can capture information at the vocabulary, syntax, semantics, etc. levels in the text. The dot product attention mechanism is a method for calculating the correlation between texts. It is used to evaluate the similarity between vectors by calculating the dot product between two vectors based on the dot product operation in the vector space.

[0050] Specifically, the special liability set, liability clause set and exemption clause set are input into the pre-trained semantic understanding model, in which the special liability set, liability clause set and exemption clause set are converted into vector representations respectively, and the dot product attention mechanism is used to calculate the difference between the vector representation of the special liability set and the vector representation of the liability clause set, and the intersection between the vector representation of the special liability set and the vector representation of the exemption clause set. Among them, the difference includes the liability clauses that exist in the special liability set but not in the liability clause set; the intersection includes the liability clauses that exist in both the special liability set and the exemption clause set.

[0051] In some optional implementations, the above-mentioned inputting the special liability set, liability clause set and exemption clause set into the pre-trained semantic understanding model, and using the dot product attention mechanism to calculate the difference between the special liability set and the liability clause set, and the intersection of the special liability set and the exemption clause set, may include the following steps:

[0052] Inputting the special liability set, liability clause set and exemption clause set into a pre-trained semantic understanding model, performing feature extraction on the special liability set, liability clause set and exemption clause set respectively, and obtaining a special liability feature vector, a liability clause feature vector and an exemption clause feature vector;

[0053] Specifically, the special liability set, liability clause set and exemption clause set are input into the pre-trained semantic understanding model. Optionally, before inputting the data into the pre-trained semantic understanding model, the special liability set, liability clause set and exemption clause set can also be pre-processed to obtain the pre-processed special liability set, liability clause set and exemption clause set. Using the pre-trained semantic understanding model, feature extraction is performed on the pre-processed special liability set, liability clause set and exemption clause set, respectively, to obtain the special liability feature vector, liability clause feature vector and exemption clause feature vector respectively.

[0054] A dot product attention mechanism is used to calculate a first dot product between the special liability feature vector and the liability clause feature vector, a first similarity is determined according to the first dot product, special liability feature vectors whose first similarity is less than a first preset threshold are selected, and a difference set between the special liability set and the liability clause set is determined;

[0055] Specifically, in the model, the dot product attention mechanism is used to calculate the first dot product between the special liability feature vector and the liability clause feature vector. According to the calculated first dot product, the first similarity is calculated by normalization or softmax function. A preset first preset threshold is obtained, the first similarity is compared with the first preset threshold, and the special liability feature vector whose first similarity is less than the first preset threshold is selected, and the selected special liability feature vector is determined as the difference set between the special liability set and the liability clause set.

[0056] A dot product attention mechanism is used to calculate the second dot product between the special liability feature vector and the disclaimer feature vector, a second similarity is determined based on the second dot product, special liability feature vectors whose second similarity is greater than a second preset threshold are selected, and the intersection of the special liability set and the disclaimer set is determined.

[0057] Specifically, in the model, the dot product attention mechanism is used to calculate the second dot product between the special liability feature vector and the exemption clause feature vector. According to the calculated second dot product, the second similarity is calculated by normalization or softmax function. A preset second preset threshold is obtained, the second similarity is compared with the second preset threshold, and the special liability feature vector whose second similarity is greater than the second preset threshold is selected, and the selected special liability feature vector is determined as the intersection of the special liability set and the exemption clause set.

[0058] The embodiment of the present application can efficiently extract the feature vectors of each of the special liability set, the liability clause set and the exemption clause set by inputting them into a pre-trained semantic understanding model; by adopting the dot product attention mechanism to calculate the dot product between the feature vectors, it can capture the subtle differences between the feature vectors, thereby achieving accurate measurement of similarity and significantly improving calculation efficiency and accuracy; by setting a preset threshold and selecting feature vectors according to the size of the similarity, it can accurately determine the difference between the special liability set and the liability clause set, as well as the intersection between the special liability set and the exemption clause set, thereby avoiding semantic misunderstandings and omissions.

[0059] In some optional implementations, before inputting the special liability set, liability clause set and exemption clause set into the pre-trained semantic understanding model, the following steps may also be included:

[0060] Acquire a pre-collected training data set, wherein the training data set includes a historical special liability set, a historical liability clause set, and a historical exemption clause set;

[0061] Input the training data set into the initial semantic understanding model, perform feature extraction on the historical special liability set, the historical liability clause set and the historical exemption clause set, respectively, and obtain the historical special liability feature vector, the historical liability clause feature vector and the historical exemption clause feature vector, wherein the initial semantic understanding model is constructed using the bidirectional encoder representation BERT model;

[0062] A dot product attention mechanism is used to calculate a third dot product between the historical special liability feature vector and the historical liability clause feature vector, a third similarity is determined according to the third dot product, historical special liability feature vectors whose third similarity is less than a third preset threshold are selected, and a training difference set between the historical special liability set and the historical liability clause set is determined;

[0063] A dot product attention mechanism is used to calculate a fourth dot product between the historical special liability feature vector and the historical exemption clause feature vector, a fourth similarity is determined according to the fourth dot product, and the historical special liability feature vector whose fourth similarity is greater than a fourth preset threshold is selected to determine a training intersection of the historical special liability set and the historical exemption clause set;

[0064] Using a preset loss function, calculate a first loss value between the historical difference set label and the training difference set, and a second loss value between the historical intersection set label and the training intersection set;

[0065] According to the first loss value and the second loss value, the model parameters of the initial semantic understanding model are updated, and when the model meets the convergence conditions, a pre-trained semantic understanding model is obtained.

[0066] Specifically, the initial semantic understanding model is built based on the bidirectional encoder BERT model, which has a dot product attention mechanism built in. The dot product attention mechanism can accurately calculate the third dot product between the historical special liability feature vector and the historical liability clause feature vector, and the fourth dot product between the historical special liability feature vector and the historical exemption clause feature vector by calculating the dot product between feature vectors at different positions.

[0067] More specifically, the process of model training for the initial semantic understanding model is as follows: first, a pre-collected training data set is obtained, and the training data set includes a historical special liability set, a historical liability clause set, and a historical exemption clause set. Then, the pre-collected training data set is input into the initial semantic understanding model, and the historical special liability set, the historical liability clause set, and the historical exemption clause set are respectively feature extracted through the initial semantic understanding model to obtain a historical special liability feature vector, a historical liability clause feature vector, and a historical exemption clause feature vector. Next, the dot product attention mechanism is used to calculate the third dot product between the historical special liability feature vector and the historical liability clause feature vector, and the third similarity is determined based on the calculated third dot product, and a pre-set third preset threshold is obtained, and the third similarity is compared with the third preset threshold. The historical special liability feature vector whose third similarity is less than the third preset threshold is selected to determine the training difference set between the historical special liability set and the historical liability clause set.

[0068] In addition, a dot product attention mechanism is used to calculate the fourth dot product between the historical special liability feature vector and the historical exemption clause feature vector, a fourth similarity is determined based on the calculated fourth dot product, a pre-set fourth preset threshold is obtained, the fourth similarity is compared with the fourth preset threshold, and the historical special liability feature vector whose fourth similarity is greater than the fourth preset threshold is selected to determine the training intersection of the historical special liability set and the historical exemption clause set.

[0069] Then, a preset loss function, a historical difference set label, and a historical intersection set label are obtained. The preset loss function is used to calculate the first loss value between the historical difference set label and the training difference set, and the second loss value between the historical intersection set label and the training intersection set is calculated. According to the calculated first loss value and second loss value, the model parameters of the initial semantic understanding model are updated, and the model training process is continuously iterated until the model meets the convergence condition, and the model training process is stopped to obtain a pre-trained semantic understanding model.

[0070] Step S204, determining the extended liability scope corresponding to the special contract text according to the difference set, and determining the exemption liability scope corresponding to the special contract text according to the intersection set;

[0071] In this embodiment, the extended liability scope refers to the additional coverage added to the special contract text relative to the insurance clause text; the exempted liability scope refers to the specific scope listed in the special contract text for which the insurance company does not bear the liability for compensation. The extended liability scope corresponding to the special contract text is determined based on the difference between the special liability set and the liability clause set; the exempted liability scope corresponding to the special contract text is determined based on the intersection of the special liability set and the exempted liability clause set.

[0072] Step S205, determining the risk level of the special liability clauses included in each liability type according to the liability types of the extended liability scope and the exempted liability scope, and determining the prompt content of the special liability clauses according to the risk level.

[0073] In this embodiment, the liability type to which each special liability clause in the extended liability scope and the exemption scope belongs is determined, wherein the liability type may include but is not limited to personal injury liability, property damage liability, medical expense liability, legal liability and other specific liability types. Different risk assessment rules are set corresponding to different liability types. For the special liability clauses contained in each liability type, the risk level of each special liability clause is evaluated one by one according to the risk assessment rules corresponding to the liability type. The risk level can be divided into two levels of low risk and high risk, or a more detailed division can be multiple levels such as low risk, lower risk, medium risk, higher risk, and high risk. According to the risk level assessed by each special liability clause, the prompt content of each special liability clause is determined accordingly, and the prompt content may include but is not limited to the underwriter review prompt information, highlight prompt information, transfer review prompt information, etc.

[0074] In some optional implementations, the above-mentioned determination of the risk level of the special liability clauses included in each liability type according to the liability types of the extended liability scope and the exempted liability scope may include the following steps:

[0075] Determine the liability type of each special liability clause included in the extended liability scope and the exemption scope respectively;

[0076] Specifically, the extended scope of liability and the scope of exemption of liability may include one or more special liability clauses. According to the characteristics of each liability type, the liability type of each special liability clause included in the extended scope of liability and the scope of exemption of liability is determined respectively. Among them, the liability type may include but is not limited to personal injury liability, property damage liability, medical expense liability, legal liability and other special liability types.

[0077] Obtaining preset risk assessment rules according to the liability type;

[0078] Specifically, different risk assessment rules are set in advance according to different liability types, and the risk assessment rules are stored in a preset risk assessment rule library. The various special liability clauses included in the extended liability scope and the exempted liability scope are classified and aggregated according to the liability type to obtain the special liability clauses included in each liability type. For the special liability clauses included in each liability type, the corresponding risk assessment rules are obtained from the preset risk assessment rule library according to the liability type.

[0079] The risk assessment rules are used to conduct a risk assessment on the special liability clauses to obtain the risk level corresponding to the special liability clauses.

[0080] Specifically, for the special liability clauses contained in each liability type, the risk assessment rules corresponding to the liability type are used to assess the risk level of each special liability clause under the liability type one by one. The risk level can be divided into two levels: low risk and high risk, or a more detailed division can be low risk, lower risk, medium risk, higher risk, high risk and other levels.

[0081] The embodiment of the present application can accurately identify the key risk points in the special contract text by respectively determining the liability types of each special liability clause included in the extended liability scope and the exemption scope, which may involve personal injury, property loss, medical expenses, legal liability, etc. Obtaining the preset risk assessment rules according to the liability type and using the rules to conduct risk assessment on the special liability clause can greatly improve the accuracy and objectivity of the risk assessment.

[0082] In some optional implementations, the risk assessment of the special liability clauses using the risk assessment rules to obtain the risk level corresponding to the special liability clauses may include the following steps:

[0083] According to the risk assessment rules, the number of liability clauses included in the target liability type is obtained;

[0084] Specifically, the target responsibility type is determined, and the number of responsibility clauses included in the target responsibility type is obtained according to the risk assessment rule corresponding to the target responsibility type, wherein the risk assessment rule may be a quantity-based assessment rule.

[0085] If the number of liability clauses is greater than a preset threshold, the risk level of the special liability clauses included in the target liability type is determined to be a high risk level;

[0086] If the number of liability clauses is less than or equal to a preset threshold, the risk level of the special liability clauses included in the target liability type is determined to be a low risk level.

[0087] Specifically, a preset quantity threshold is extracted from the quantity-based assessment rule, and the number of liability clauses obtained is compared with the preset quantity threshold. If the number of liability clauses is greater than the preset quantity threshold, it indicates that the special liability clauses under the target liability type are relatively complex and may involve a higher compensation risk, and the risk level of the special liability clauses contained in the target liability type is determined to be a high risk level. If the number of liability clauses is less than or equal to the preset quantity threshold, it indicates that the special liability clauses under the target liability type are relatively simple and the compensation risk is relatively low, and the risk level of the special liability clauses contained in the target liability type is determined to be a low risk level.

[0088] The embodiment of the present application counts the number of special liability clauses under the target liability type and compares it with a preset quantity threshold. It can conduct preliminary risk screening on a large number of liability clauses in a short period of time, quickly classify the liability types into high-risk or low-risk levels, and greatly improve the efficiency of risk assessment.

[0089] In some optional implementations, the above-mentioned determining the prompt content of the special liability clause according to the risk level may include the following steps:

[0090] If the risk level is a low risk level, a highlighted prompt message is generated according to the special liability clause corresponding to the low risk level;

[0091] Specifically, if the risk level of the special liability clause is low risk, the low risk level prompt content generation process is executed, and a highlighted prompt message is generated according to the special liability clause corresponding to the low risk level. The highlighting can be performed by color, bold font or other visual prominence to attract the attention of users or auditors.

[0092] If the risk level is a high risk level, a transfer review prompt message is generated according to the special liability clause corresponding to the high risk level.

[0093] Specifically, if the risk level of the special liability clause is high, the high risk level prompt content generation process is executed, and the transfer review prompt information is generated according to the special liability clause corresponding to the high risk level. The transfer review prompt information contains more detailed clause content, potential risk points, the review department recommended for transfer, the review process or precautions, etc., to ensure that the high-risk clauses can be assigned to more professional departments for review.

[0094] The embodiment of the present application can quickly identify low-risk and high-risk clauses by generating different forms of prompt information for special liability clauses according to risk levels. The highlighting of low-risk clauses allows auditors to quickly browse and confirm their risk controllability, while the transfer review prompts for high-risk clauses ensure that these clauses can be deeply reviewed and evaluated by professional teams. This differentiated processing method greatly improves the efficiency of risk identification.

[0095] The embodiment of the present application uses the above scheme, specifically by adopting the preset liability set extraction rules, to accurately extract the liability set from the special text and the insurance clause text, including the special liability set, the liability clause set and the exemption clause set; by adopting the pre-trained semantic understanding model, the semantic information in the special text and the insurance clause text can be deeply understood, and the dot product attention mechanism is used to calculate the difference between the special liability set and the liability clause set, as well as the intersection of the special liability set and the exemption clause set, so as to accurately identify the extended liability scope and the exemption liability scope corresponding to the special text, and by abstracting the semantic problem into a mathematical problem and processing the liability definition in a set manner, compared with the traditional semantic comparison reasoning, not only the working mechanism is clear, easy to trace back and explain, but also the accuracy and efficiency of recognition are improved; by determining the risk level of the special liability clauses contained in each liability type according to the liability type of the extended liability scope and the exemption liability scope, and determining the prompt content of the special liability clause according to the risk level, the business process is made more scientific and reasonable, and the hierarchical management and control of risks can be achieved, and the efficiency and effect of business processing can be improved.

[0096] It should be emphasized that in order to further ensure the privacy and security of the above-mentioned special text to be identified, insurance clause text, pre-trained semantic understanding model and other information as well as other relevant data, the above-mentioned special text to be identified, insurance clause text, pre-trained semantic understanding model and other information as well as other relevant data can also be stored in a blockchain node.

[0097] The blockchain referred to in this application is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.

[0098] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0099] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0100] Further references Figure 3 , as a response to the above Figure 2 The present application provides an embodiment of a device for identifying insurance policy risks. Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0101] like Figure 3 As shown, the insurance policy risk identification device 400 described in this embodiment includes: an acquisition module 401, an extraction module 402, a calculation module 403, a determination module 404 and a classification module 405. Among them:

[0102] The acquisition module 401 is used to acquire the special contract text to be identified and the insurance clause text corresponding to the special contract text;

[0103] Extraction module 402, used to extract a special liability set from the special contract text, and extract a liability clause set and a disclaimer clause set from the insurance clause text by using a preset liability set extraction rule, wherein the special liability set, liability clause set and disclaimer clause set are all composed of minimum semantic units;

[0104] A calculation module 403 is used to input the special liability set, the liability clause set and the exemption clause set into a pre-trained semantic understanding model, and use a dot product attention mechanism to calculate the difference between the special liability set and the liability clause set, and the intersection between the special liability set and the exemption clause set;

[0105] A determination module 404 is used to determine the extended liability scope corresponding to the special contract text according to the difference set, and determine the exemption liability scope corresponding to the special contract text according to the intersection set;

[0106] The grading module 405 is used to determine the risk level of the special liability clauses included in each liability type according to the liability types of the extended liability scope and the exempted liability scope, and determine the prompt content of the special liability clauses according to the risk level.

[0107] In this embodiment, the acquisition module 401 is used to obtain the special text to be identified and the insurance clause text corresponding to the special text. The special text refers to the special liability clauses agreed upon by the two parties (the insurance company and the insured) outside the main contract based on specific needs or circumstances during the signing process of the insurance contract. The special liability clauses may involve specific coverage, liability limitations, exemption conditions, claims conditions, etc. The insurance clause text is the core content of the insurance contract, which stipulates the rights and obligations between the insurance company and the insured, including insurance liability, exemption clauses, claims process, etc. The insurance clause text may include the main insurance clause text and the supplementary insurance clause text, wherein the main insurance clause text is the clause part that stipulates the basic content of the insurance contract and the basic rights and obligations of both parties; the supplementary insurance clause text is the clause part that supplements or expands the main insurance clause text.

[0108] Specifically, the extraction module 402 is used to pre-set a liability set extraction rule, which is formulated according to the characteristics of the insurance clauses and may include a special liability extraction rule, a liability clause extraction rule, an exemption clause extraction rule, etc. The preset liability set extraction rule is obtained, and the special text is analyzed sentence by sentence using the liability set extraction rule, split into the smallest semantic unit, and the special liability set is extracted; the insurance clause text is analyzed sentence by sentence using the liability set extraction rule, split into the smallest semantic unit, and the liability clause set and the exemption clause set are extracted.

[0109] Specifically, the semantic understanding model is a machine learning model used to understand and process text semantic information. The model is built based on deep learning technology and can capture information at the vocabulary, syntax, semantics and other levels in the text. The dot product attention mechanism is a method for calculating the correlation between texts. It is used to evaluate the similarity between vectors by calculating the dot product between two vectors based on the dot product operation in the vector space.

[0110] Specifically, the calculation module 403 is used to input the special liability set, the liability clause set and the exemption clause set into the pre-trained semantic understanding model, in which the special liability set, the liability clause set and the exemption clause set are converted into vector representations respectively, and the dot product attention mechanism is used to calculate the difference between the vector representation of the special liability set and the vector representation of the liability clause set, and the intersection between the vector representation of the special liability set and the vector representation of the exemption clause set. Among them, the difference includes the liability clauses that exist in the special liability set but not in the liability clause set; the intersection includes the liability clauses that exist in both the special liability set and the exemption clause set.

[0111] Specifically, the extended liability scope refers to the additional coverage added to the special contract relative to the insurance clause text; the exempted liability scope refers to the specific scope listed in the special contract for which the insurance company does not bear the liability for compensation. The determination module 404 is used to determine the extended liability scope corresponding to the special contract based on the difference between the special liability set and the liability clause set; and to determine the exempted liability scope corresponding to the special contract based on the intersection of the special liability set and the exempted liability clause set.

[0112] Specifically, the classification module 405 is used to determine the liability type to which each special liability clause in the extended liability scope and the exemption scope belongs, wherein the liability type may include but is not limited to personal injury liability, property damage liability, medical expense liability, legal liability and other specific liability types. Different risk assessment rules are set corresponding to different liability types. For the special liability clauses contained in each liability type, the risk level of each special liability clause is evaluated one by one according to the risk assessment rules corresponding to the liability type. The risk level can be divided into two levels of low risk and high risk, or a more detailed division can be multiple levels such as low risk, lower risk, medium risk, higher risk, and high risk. According to the risk level assessed by each special liability clause, the prompt content of each special liability clause is determined accordingly, and the prompt content may include but is not limited to the underwriter review prompt information, highlight prompt information, transfer review prompt information, etc.

[0113] The insurance policy risk identification device 400 of the embodiment of the present application can accurately extract the liability set from the special text and the insurance clause text by adopting the preset liability set extraction rule, including the special liability set, the liability clause set and the exemption clause set; by adopting the pre-trained semantic understanding model, it can deeply understand the semantic information in the special text and the insurance clause text, and use the dot product attention mechanism to calculate the difference between the special liability set and the liability clause set, and the intersection between the special liability set and the exemption clause set, so as to accurately identify the extended liability scope and the exemption liability scope corresponding to the special text, and by abstracting the semantic problem into a mathematical problem and processing the liability definition in a set manner, compared with the traditional semantic comparison reasoning, not only the working mechanism is clear, easy to trace back and explain, but also the accuracy and efficiency of identification are improved; by determining the risk level of the special liability clauses contained in each liability type according to the liability type of the extended liability scope and the exemption liability scope, and determining the prompt content of the special liability clause according to the risk level, the business process is made more scientific and reasonable, and the hierarchical control of risks can be realized, and the efficiency and effect of business processing can be improved.

[0114] In some optional implementations of this embodiment, the extraction module 402 may include a template acquisition submodule, a first matching submodule, and a second matching submodule.

[0115] The template acquisition submodule is used to acquire a preset responsibility description template, wherein the responsibility description template includes a responsibility tag related to the responsibility content, wherein the responsibility tag is set according to a preset responsibility set extraction rule, and wherein the responsibility tag includes a special responsibility tag, a responsibility scope tag, and an exemption scope tag;

[0116] A first matching submodule is used to use the responsibility description template to extract the minimum semantic unit matching the special responsibility label from the special text as a special responsibility set;

[0117] The second matching submodule is used to adopt the liability description template to extract the minimum semantic unit that matches the liability scope label from the insurance clause text as a liability clause set, and to extract the minimum semantic unit that matches the exemption scope label from the insurance clause text as a exemption clause set.

[0118] Specifically, the template acquisition submodule is used to obtain a preset liability description template. The liability description template may include a liability tag related to the liability content. The liability tag is set according to a preset liability set extraction rule. The liability set extraction rule may include a special liability extraction rule, a liability clause extraction rule, and an exemption clause extraction rule. The liability tag may include a special liability tag, a liability scope tag, and an exemption scope tag. Each liability tag corresponds to a specific liability content, such as the liability object (insured, policyholder), liability event (accidental injury, disease diagnosis), compensation amount, compensation ratio, compensation conditions, and the like.

[0119] Specifically, the first matching submodule is used to apply the acquired responsibility description template to the special text, split the special text into multiple minimum semantic units, and compare the special responsibility label in the responsibility description template with each of the split minimum semantic units one by one, and extract the minimum semantic unit that matches the special responsibility label as the special responsibility set.

[0120] Specifically, the second matching submodule is used to apply the acquired liability description template to the insurance clause text, split the insurance clause text into multiple minimum semantic units, and compare the liability scope label in the liability description template with each of the split minimum semantic units one by one, and extract the minimum semantic units that match the liability scope label as the liability clause set. Compare the exemption scope label in the liability description template with each of the split minimum semantic units one by one, and extract the minimum semantic units that match the exemption scope label as the exemption clause set.

[0121] The extraction module 402 of the embodiment of the present application sets a liability description template according to the liability set extraction rules, compares the liability labels and the various text contents in the liability description template, and can accurately identify the smallest semantic unit that matches the liability label, and efficiently extract key liability information from the special text and insurance clause text, thereby avoiding semantic misunderstandings and omissions.

[0122] In some optional implementations of this embodiment, the calculation module 403 may include a feature extraction submodule, a first calculation submodule, and a second calculation submodule.

[0123] A feature extraction submodule, for inputting the special liability set, liability clause set and exemption clause set into a pre-trained semantic understanding model, and performing feature extraction on the special liability set, liability clause set and exemption clause set, respectively, to obtain a special liability feature vector, a liability clause feature vector and an exemption clause feature vector;

[0124] A first calculation submodule, configured to calculate a first dot product between the special liability feature vector and the liability clause feature vector using a dot product attention mechanism, determine a first similarity based on the first dot product, select a special liability feature vector whose first similarity is less than a first preset threshold, and determine a difference set between the special liability set and the liability clause set;

[0125] The second calculation submodule is used to calculate the second dot product between the special liability feature vector and the disclaimer feature vector using the dot product attention mechanism, determine the second similarity based on the second dot product, select the special liability feature vector whose second similarity is greater than a second preset threshold, and determine the intersection of the special liability set and the disclaimer set.

[0126] Specifically, the feature extraction submodule is used to input the special liability set, the liability clause set and the exemption clause set into the pre-trained semantic understanding model. Optionally, before inputting the data into the pre-trained semantic understanding model, the special liability set, the liability clause set and the exemption clause set may be pre-processed to obtain the pre-processed special liability set, the liability clause set and the exemption clause set. Using the pre-trained semantic understanding model, feature extraction is performed on the pre-processed special liability set, the liability clause set and the exemption clause set, respectively, to obtain the special liability feature vector, the liability clause feature vector and the exemption clause feature vector respectively.

[0127] Specifically, the first calculation submodule is used to calculate the first dot product between the special liability feature vector and the liability clause feature vector in the model by using the dot product attention mechanism. According to the calculated first dot product, the first similarity is calculated by normalization or softmax function. A preset first preset threshold is obtained, the first similarity is compared with the first preset threshold, the special liability feature vector whose first similarity is less than the first preset threshold is selected, and the selected special liability feature vector is determined as the difference set between the special liability set and the liability clause set.

[0128] Specifically, the second calculation submodule is used to calculate the second dot product between the special liability feature vector and the exemption clause feature vector in the model by using the dot product attention mechanism. According to the calculated second dot product, the second similarity is calculated by normalization or softmax function. A preset second preset threshold is obtained, the second similarity is compared with the second preset threshold, the special liability feature vector whose second similarity is greater than the second preset threshold is selected, and the selected special liability feature vector is determined as the intersection of the special liability set and the exemption clause set.

[0129] The calculation module 403 of the embodiment of the present application can efficiently extract the respective feature vectors by inputting the special liability set, the liability clause set and the exemption clause set into the pre-trained semantic understanding model; by adopting the dot product attention mechanism to calculate the dot product between the feature vectors, it can capture the subtle differences between the feature vectors, thereby achieving accurate measurement of similarity and significantly improving calculation efficiency and accuracy; by setting a preset threshold and selecting feature vectors according to the size of the similarity, it can accurately determine the difference between the special liability set and the liability clause set, as well as the intersection of the special liability set and the exemption clause set, thereby avoiding semantic misunderstandings and omissions.

[0130] In some optional implementations of this embodiment, the insurance policy risk identification device 400 may further include a model training module. The model training module is used to obtain a pre-collected training data set, wherein the training data set includes a historical special liability set, a historical liability clause set, and a historical exemption clause set;

[0131] Input the training data set into the initial semantic understanding model, perform feature extraction on the historical special liability set, the historical liability clause set and the historical exemption clause set, respectively, and obtain the historical special liability feature vector, the historical liability clause feature vector and the historical exemption clause feature vector, wherein the initial semantic understanding model is constructed using the bidirectional encoder representation BERT model;

[0132] A dot product attention mechanism is used to calculate a third dot product between the historical special liability feature vector and the historical liability clause feature vector, a third similarity is determined according to the third dot product, historical special liability feature vectors whose third similarity is less than a third preset threshold are selected, and a training difference set between the historical special liability set and the historical liability clause set is determined;

[0133] A dot product attention mechanism is used to calculate a fourth dot product between the historical special liability feature vector and the historical exemption clause feature vector, a fourth similarity is determined according to the fourth dot product, and the historical special liability feature vector whose fourth similarity is greater than a fourth preset threshold is selected to determine a training intersection of the historical special liability set and the historical exemption clause set;

[0134] Using a preset loss function, calculate a first loss value between the historical difference set label and the training difference set, and a second loss value between the historical intersection set label and the training intersection set;

[0135] According to the first loss value and the second loss value, the model parameters of the initial semantic understanding model are updated, and when the model meets the convergence conditions, a pre-trained semantic understanding model is obtained.

[0136] Specifically, the initial semantic understanding model is built based on the bidirectional encoder BERT model, which has a dot product attention mechanism built in. The dot product attention mechanism can accurately calculate the third dot product between the historical special liability feature vector and the historical liability clause feature vector, and the fourth dot product between the historical special liability feature vector and the historical exemption clause feature vector by calculating the dot product between feature vectors at different positions.

[0137] More specifically, the process of model training for the initial semantic understanding model is as follows: first, a pre-collected training data set is obtained, and the training data set includes a historical special liability set, a historical liability clause set, and a historical exemption clause set. Then, the pre-collected training data set is input into the initial semantic understanding model, and the historical special liability set, the historical liability clause set, and the historical exemption clause set are respectively feature extracted through the initial semantic understanding model to obtain a historical special liability feature vector, a historical liability clause feature vector, and a historical exemption clause feature vector. Next, the dot product attention mechanism is used to calculate the third dot product between the historical special liability feature vector and the historical liability clause feature vector, and the third similarity is determined based on the calculated third dot product, and a pre-set third preset threshold is obtained, and the third similarity is compared with the third preset threshold. The historical special liability feature vector whose third similarity is less than the third preset threshold is selected to determine the training difference set between the historical special liability set and the historical liability clause set.

[0138] In addition, a dot product attention mechanism is used to calculate the fourth dot product between the historical special liability feature vector and the historical exemption clause feature vector, a fourth similarity is determined based on the calculated fourth dot product, a pre-set fourth preset threshold is obtained, the fourth similarity is compared with the fourth preset threshold, and the historical special liability feature vector whose fourth similarity is greater than the fourth preset threshold is selected to determine the training intersection of the historical special liability set and the historical exemption clause set.

[0139] Then, a preset loss function, a historical difference set label, and a historical intersection set label are obtained. The preset loss function is used to calculate the first loss value between the historical difference set label and the training difference set, and the second loss value between the historical intersection set label and the training intersection set is calculated. According to the calculated first loss value and second loss value, the model parameters of the initial semantic understanding model are updated, and the model training process is continuously iterated until the model meets the convergence condition, and the model training process is stopped to obtain a pre-trained semantic understanding model.

[0140] In some optional implementations of this embodiment, the classification module 405 may include a type determination submodule, a rule acquisition submodule, and a risk assessment submodule.

[0141] A type determination submodule, used to respectively determine the liability type of each special liability clause included in the extended liability scope and the exemption scope;

[0142] A rule acquisition submodule, used to acquire preset risk assessment rules according to the responsibility type;

[0143] The risk assessment submodule is used to adopt the risk assessment rules to conduct risk assessment on the special liability clauses to obtain the risk level corresponding to the special liability clauses.

[0144] Specifically, the extended scope of liability and the scope of exemption of liability may include one or more special liability clauses. The type determination submodule is used to determine the liability type of each special liability clause included in the extended scope of liability and the scope of exemption of liability according to the characteristics of each liability type. Among them, the liability type may include but is not limited to personal injury liability, property damage liability, medical expense liability, legal liability and other special liability types.

[0145] Specifically, the rule acquisition submodule is used to pre-set different risk assessment rules according to different liability types, and store the risk assessment rules in a preset risk assessment rule library. The various special liability clauses included in the extended liability scope and the exemption scope are classified and aggregated according to the liability type to obtain the special liability clauses included in each liability type. For the special liability clauses included in each liability type, the corresponding risk assessment rules are obtained from the preset risk assessment rule library according to the liability type.

[0146] Specifically, the risk assessment submodule is used to assess the risk level of each special liability clause contained in each liability type by using the risk assessment rules corresponding to the liability type. The risk level can be divided into two levels: low risk and high risk, or a more detailed division can be low risk, lower risk, medium risk, higher risk, high risk and other levels.

[0147] The classification module 405 of the embodiment of the present application can accurately identify the key risk points in the special contract text by respectively determining the liability types of each special liability clause included in the extended liability scope and the exemption scope. These risk points may involve personal injury, property loss, medical expenses, legal liability, etc. Obtaining the preset risk assessment rules according to the liability type and using the rules to conduct risk assessment on the special liability clause can greatly improve the accuracy and objectivity of the risk assessment.

[0148] In some optional implementations of this embodiment, the risk assessment submodule may include a quantity acquisition unit, a first determination unit, and a second determination unit.

[0149] A quantity acquisition unit, used to acquire the quantity of liability clauses included in the target liability type according to the risk assessment rule;

[0150] A first determination unit is used to determine the risk level of the special liability clauses included in the target liability type as a high risk level if the number of the liability clauses is greater than a preset number threshold;

[0151] The second determination unit is used to determine the risk level of the special liability clauses included in the target liability type as a low risk level if the number of the liability clauses is less than or equal to a preset number threshold.

[0152] Specifically, the quantity acquisition unit is used to determine the target responsibility type, and obtain the number of responsibility clauses included in the target responsibility type according to the risk assessment rule corresponding to the target responsibility type, wherein the risk assessment rule can be a quantity-based assessment rule.

[0153] Specifically, the first determination unit is used to extract a preset quantity threshold from the quantity-based assessment rule, and compare the acquired number of liability clauses with the preset quantity threshold. If the number of liability clauses is greater than the preset quantity threshold, it indicates that the special liability clauses under the target liability type are relatively complex and may involve a higher compensation risk, and the risk level of the special liability clauses contained in the target liability type is determined to be a high risk level. The second determination unit is used to determine the risk level of the special liability clauses contained in the target liability type as a low risk level if the number of liability clauses is less than or equal to the preset quantity threshold, indicating that the special liability clauses under the target liability type are relatively simple and the compensation risk is relatively low.

[0154] The risk assessment submodule of the embodiment of the present application can conduct preliminary risk screening on a large number of liability clauses in a short period of time by counting the number of special liability clauses under the target liability type and comparing it with a preset quantity threshold, and quickly classify the liability types into high-risk or low-risk levels, thereby greatly improving the efficiency of risk assessment.

[0155] In some optional implementations of this embodiment, the grading module 405 may include a first prompting submodule and a second prompting submodule.

[0156] A first prompt submodule, for generating a highlighted prompt message according to the special liability clause corresponding to the low risk level if the risk level is a low risk level;

[0157] The second prompt submodule is used to generate transfer review prompt information according to the special liability clause corresponding to the high risk level if the risk level is a high risk level.

[0158] Specifically, the first prompt submodule is used to execute the low risk level prompt content generation process if the risk level of the special liability clause is a low risk level, and generate a highlighted prompt information according to the special liability clause corresponding to the low risk level. The highlighted prompt can be provided by color, bold font or other visual prominence to attract the attention of users or auditors.

[0159] Specifically, the second prompt submodule is used to execute the high-risk level prompt content generation process if the risk level of the special liability clause is a high-risk level, and generate a transfer review prompt information according to the special liability clause corresponding to the high-risk level. The transfer review prompt information contains more detailed clause content, potential risk points, the audit department recommended for transfer, the audit process or precautions, etc., to ensure that high-risk clauses can be assigned to more professional departments for review.

[0160] The classification module 405 of the embodiment of the present application can quickly identify low-risk and high-risk clauses by generating different forms of prompt information for special liability clauses according to risk levels. The highlighting of low-risk clauses allows auditors to quickly browse and confirm their risk controllability, while the transfer review prompt of high-risk clauses ensures that these clauses can be deeply reviewed and evaluated by professional teams. This differentiated processing method greatly improves the efficiency of risk identification.

[0161] To solve the above technical problems, the present application also provides a computer device. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0162] The computer device 6 includes a memory 61, a processor 62, and a network interface 63 that are interconnected and communicated through a system bus. It should be noted that the figure only shows a computer device 6 with a memory 61, a processor 62, and a network interface 63, 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.

[0163] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0164] The memory 61 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 61 can be an internal storage unit of the computer device 6, such as a hard disk or memory of the computer device 6. In other embodiments, the memory 61 can also be an external storage device of the computer device 6, such as a plug-in hard disk equipped on the computer device 6, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. Of course, the memory 61 can also include both the internal storage unit of the computer device 6 and its external storage device. In this embodiment, the memory 61 is generally used to store the operating system and various application software installed on the computer device 6, such as computer-readable instructions of the insurance policy risk identification method, etc. In addition, the memory 61 can also be used to temporarily store various types of data that have been output or are to be output.

[0165] The processor 62 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor or other data processing chip in some embodiments. The processor 62 is generally used to control the overall operation of the computer device 6. In this embodiment, the processor 62 is used to run the computer-readable instructions stored in the memory 61 or process data, such as computer-readable instructions for running the insurance policy risk identification method.

[0166] The network interface 63 may include a wireless network interface or a wired network interface. The network interface 63 is generally used to establish a communication connection between the computer device 6 and other electronic devices.

[0167] The present application also provides another embodiment, namely, providing a computer-readable storage medium, which stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the insurance policy risk identification method as described above.

[0168] The computer device, computer-readable storage medium and computer-readable instructions thereof provided in the embodiment of the present application can accurately extract liability sets, including special liability sets, liability clause sets and exemption clause sets, from special texts and insurance clause texts by executing preset liability set extraction rules through a processor; by adopting a pre-trained semantic understanding model, the semantic information in the special texts and insurance clause texts can be deeply understood, and the difference between the special liability set and the liability clause set, as well as the intersection of the special liability set and the exemption clause set, can be calculated by using a dot product attention mechanism, so as to accurately identify the extended liability scope and the exemption liability scope corresponding to the special text; by abstracting the semantic problem into a mathematical problem and processing the liability definition in a set-based manner, compared with traditional semantic comparison reasoning, not only the working mechanism is clear, easy to trace back and explain, but also the accuracy and efficiency of recognition are improved; by determining the risk level of the special liability clauses contained in each liability type according to the liability type of the extended liability scope and the exemption liability scope, and determining the prompt content of the special liability clause according to the risk level, the business process is made more scientific and reasonable, and the hierarchical management and control of risks can be achieved, thereby improving the efficiency and effect of business processing.

[0169] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0170] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to replace some of the technical features therein with equivalents. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

[0171] The non-Company software tools or components appearing in the embodiments of this application are merely examples and do not represent actual use.

Claims

1. A method for identifying insurance policy risks, characterized in that: The steps include: Obtaining the special contract text to be identified and the insurance clause text corresponding to the special contract text; Using a preset liability set extraction rule, extract a special liability set from the special contract text, and extract a liability clause set and an exemption clause set from the insurance clause text, wherein the special liability set, the liability clause set and the exemption clause set are all composed of minimum semantic units; Input the special liability set, liability clause set and exemption clause set into the pre-trained semantic understanding model, and use the dot product attention mechanism to calculate the difference between the special liability set and the liability clause set, and the intersection between the special liability set and the exemption clause set; Determine the extended liability scope corresponding to the special contract text according to the difference set, and determine the exemption liability scope corresponding to the special contract text according to the intersection set; According to the liability types of the extended liability scope and the exempted liability scope, the risk level of the special liability clauses included in each liability type is determined, and the prompt content of the special liability clauses is determined according to the risk level.

2. The insurance policy risk identification method according to claim 1, characterized in that: The step of extracting a special liability set from the special contract text by using a preset liability set extraction rule, and extracting a liability clause set and an exemption clause set from the insurance clause text comprises: Obtaining a preset responsibility description template, wherein the responsibility description template includes a responsibility tag related to the responsibility content, wherein the responsibility tag is set according to a preset responsibility set extraction rule, and wherein the responsibility tag includes a special responsibility tag, a responsibility scope tag, and an exemption scope tag; Using the responsibility description template, extracting the minimum semantic unit matching the special responsibility label from the special text as a special responsibility set; The liability description template is used to extract the smallest semantic unit that matches the liability scope label from the insurance clause text as a liability clause set, and the smallest semantic unit that matches the exemption scope label is extracted from the insurance clause text as an exemption clause set.

3. The insurance policy risk identification method according to claim 1, characterized in that: The step of inputting the special liability set, liability clause set and exemption clause set into the pre-trained semantic understanding model, and using the dot product attention mechanism to calculate the difference between the special liability set and the liability clause set, and the intersection between the special liability set and the exemption clause set, comprises: Inputting the special liability set, liability clause set and exemption clause set into a pre-trained semantic understanding model, performing feature extraction on the special liability set, liability clause set and exemption clause set respectively, and obtaining a special liability feature vector, a liability clause feature vector and an exemption clause feature vector; A dot product attention mechanism is used to calculate a first dot product between the special liability feature vector and the liability clause feature vector, a first similarity is determined according to the first dot product, special liability feature vectors whose first similarity is less than a first preset threshold are selected, and a difference set between the special liability set and the liability clause set is determined; A dot product attention mechanism is used to calculate the second dot product between the special liability feature vector and the disclaimer feature vector, a second similarity is determined based on the second dot product, special liability feature vectors whose second similarity is greater than a second preset threshold are selected, and the intersection of the special liability set and the disclaimer set is determined.

4. The insurance policy risk identification method according to claim 1, characterized in that: Before the step of inputting the special liability set, liability clause set and exemption clause set into the pre-trained semantic understanding model, the method further includes: Acquire a pre-collected training data set, wherein the training data set includes a historical special liability set, a historical liability clause set, and a historical exemption clause set; Input the training data set into the initial semantic understanding model, perform feature extraction on the historical special liability set, the historical liability clause set and the historical exemption clause set, respectively, and obtain the historical special liability feature vector, the historical liability clause feature vector and the historical exemption clause feature vector, wherein the initial semantic understanding model is constructed using the bidirectional encoder representation BERT model; A dot product attention mechanism is used to calculate a third dot product between the historical special liability feature vector and the historical liability clause feature vector, a third similarity is determined according to the third dot product, historical special liability feature vectors whose third similarity is less than a third preset threshold are selected, and a training difference set between the historical special liability set and the historical liability clause set is determined; A dot product attention mechanism is used to calculate a fourth dot product between the historical special liability feature vector and the historical exemption clause feature vector, a fourth similarity is determined according to the fourth dot product, and the historical special liability feature vector whose fourth similarity is greater than a fourth preset threshold is selected to determine a training intersection of the historical special liability set and the historical exemption clause set; Using a preset loss function, calculate a first loss value between the historical difference set label and the training difference set, and a second loss value between the historical intersection set label and the training intersection set; According to the first loss value and the second loss value, the model parameters of the initial semantic understanding model are updated, and when the model meets the convergence conditions, a pre-trained semantic understanding model is obtained.

5. The insurance policy risk identification method according to claim 1, characterized in that: The step of determining the risk level of the special liability clauses included in each liability type according to the liability types of the extended liability scope and the exempted liability scope includes: Determine the liability type of each special liability clause included in the extended liability scope and the exemption scope respectively; Obtaining preset risk assessment rules according to the liability type; The risk assessment rules are used to conduct a risk assessment on the special liability clauses to obtain the risk level corresponding to the special liability clauses.

6. The insurance policy risk identification method according to claim 5, characterized in that: The step of using the risk assessment rule to conduct risk assessment on the special liability clause to obtain the risk level corresponding to the special liability clause includes: According to the risk assessment rules, the number of liability clauses included in the target liability type is obtained; If the number of liability clauses is greater than a preset threshold, the risk level of the special liability clauses included in the target liability type is determined to be a high risk level; If the number of liability clauses is less than or equal to a preset threshold, the risk level of the special liability clauses included in the target liability type is determined to be a low risk level.

7. The insurance policy risk identification method according to claim 6, characterized in that: The step of determining the prompt content of the special liability clause according to the risk level includes: If the risk level is a low risk level, a highlighted prompt message is generated according to the special liability clause corresponding to the low risk level; If the risk level is a high risk level, a transfer review prompt message is generated according to the special liability clause corresponding to the high risk level.

8. A device for identifying insurance policy risks, characterized in that: The device comprises: An acquisition module, used to acquire the special contract text to be identified and the insurance clause text corresponding to the special contract text; An extraction module, used to extract a special liability set from the special contract text, and extract a liability clause set and a disclaimer clause set from the insurance clause text by using a preset liability set extraction rule, wherein the special liability set, the liability clause set and the disclaimer clause set are all composed of minimum semantic units; A calculation module, used for inputting the special liability set, the liability clause set and the exemption clause set into a pre-trained semantic understanding model, and using a dot product attention mechanism to calculate the difference between the special liability set and the liability clause set, and the intersection between the special liability set and the exemption clause set; A determination module, used to determine the extended liability scope corresponding to the special contract text according to the difference set, and determine the exemption liability scope corresponding to the special contract text according to the intersection set; A grading module is used to determine the risk level of the special liability clauses included in each liability type according to the liability types of the extended liability scope and the exempted liability scope, and to determine the prompt content of the special liability clauses according to the risk level.

9. A computer device, characterized in that: It comprises a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the steps of the insurance policy risk identification method as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by a processor, the steps of the insurance policy risk identification method as described in any one of claims 1 to 7 are implemented.

Images