1. A risk management priority evaluation method facing intelligent driving artificial intelligence is characterized by comprising the following steps:

acquiring technical requirements;

determining, based on the technical requirement, an outcome of not meeting the technical requirement;

analyzing and obtaining risk elements according to the consequences;

constructing a hierarchical structure for risk element treatment priority evaluation;

determining the emergency degree weight and the importance degree weight of the hierarchical structure by adopting an expert scoring method;

determining the treatment priority of the risk elements according to the emergency degree weight of the hierarchical structure and the importance degree weight of the hierarchical structure; the governance priorities include: a first priority, a second priority, and a third priority.

2. The intelligent driving artificial intelligence oriented risk management priority assessment method according to claim 1, wherein the constructing of the hierarchical structure of risk element management priority assessment specifically comprises:

constructing the hierarchical structure by taking the artificial intelligence risk management priority facing intelligent driving as a target layer, taking the risk category as a criterion layer and taking the risk elements as an index layer; the risk elements include algorithmic risk elements and data risk elements; the risk categories include algorithmic risk and data risk.

3. The risk management priority evaluation method for the artificial intelligence of intelligent driving according to claim 2, wherein the determining the emergency degree weight and the importance degree weight of the hierarchical structure by using an expert scoring method specifically comprises:

determining the importance degree weight of the criterion layer by adopting an expert scoring method;

determining an importance weight of the criterion layer relative to the target layer based on the importance weight of the criterion layer;

determining the importance degree weight of the index layer relative to the criterion layer by adopting an expert scoring method;

determining a importance level weight of the index layer relative to the target layer based on the importance level weight of the criterion layer relative to the target layer and the importance level weight of the index layer relative to the criterion layer;

carrying out standardization processing on the importance degree weight of the index layer relative to the target layer to obtain the importance degree weight of the hierarchical structure; the importance degree weight of the hierarchical structure is the importance degree weight of the risk elements relative to the target layer;

determining the emergency degree weight of the criterion layer by adopting an expert scoring method;

determining the emergency degree weight of the index layer relative to the standard layer by adopting an expert scoring method based on the emergency degree weight of the standard layer;

determining an urgency weight of the index layer relative to the target layer based on the urgency weight of the criterion layer and the urgency weight of the index layer relative to the criterion layer;

carrying out standardization processing on the emergency degree weight of the index layer relative to the target layer to obtain the emergency degree weight of the hierarchical structure; the urgency weight of the hierarchy is the urgency weight of the risk element relative to the target tier.

4. The risk management priority evaluation method for intelligent driving artificial intelligence according to claim 1, wherein the determining the management priority of the risk elements according to the emergency degree weight of the hierarchical structure and the importance degree weight of the hierarchical structure specifically comprises:

determining an importance level of the risk element and an urgency level of the risk element according to the importance weight of the hierarchical structure and the importance weight of the hierarchical structure; the importance level comprises importance and next importance; the urgency class includes urgent and sub-urgent;

constructing an important emergency matrix according to the importance degree grade of the risk element and the emergency degree grade of the risk element;

and determining the treatment priority of the risk elements according to the important emergency matrix.

5. The method for assessing the priority of risk management facing the artificial intelligence of intelligent driving according to claim 4, wherein the determining the importance level of the risk element and the urgency level of the risk element according to the importance weight of the hierarchical structure and the urgency weight of the hierarchical structure specifically comprises:

acquiring a first preset value and a second preset value;

when the importance degree weight of the hierarchical structure is larger than a first preset value, the importance degree grade of the risk elements is important;

when the importance degree weight of the hierarchical structure is less than or equal to the first preset value, the importance degree grade of the risk elements is secondary importance;

when the emergency degree weight of the hierarchical structure is greater than a second preset value, the emergency degree grade of the risk elements is emergency;

and when the emergency degree weight of the hierarchical structure is less than or equal to the second preset value, the emergency degree grade of the risk elements is sub-emergency.

6. The utility model provides a risk management priority evaluation system towards intelligent driving artificial intelligence which characterized in that includes:

the technical requirement acquisition module is used for acquiring technical requirements;

an outcome determination module to determine an outcome that does not meet the specification based on the specification;

the risk element analysis module is used for analyzing and obtaining a risk element according to the result;

the hierarchical structure construction module is used for constructing a hierarchical structure for risk element treatment priority evaluation;

the weight determining module is used for determining the emergency degree weight and the importance degree weight of the hierarchical structure by adopting an expert scoring method;

the treatment priority determining module is used for determining the treatment priority of the risk elements according to the emergency degree weight of the hierarchical structure and the importance degree weight of the hierarchical structure; the governance priorities include: a first priority, a second priority, and a third priority.

7. The risk management priority evaluation system for intelligent driving artificial intelligence of claim 6, wherein the hierarchical structure building module specifically comprises:

the hierarchical structure building unit is used for building the hierarchical structure by taking the artificial intelligence risk management priority facing intelligent driving as a target layer, taking the risk category as a criterion layer and taking the risk elements as an index layer; the risk elements include algorithmic risk elements and data risk elements; the risk categories include algorithmic risk and data risk.

8. The risk management priority evaluation system for intelligent driving artificial intelligence of claim 7, wherein the weight determination module specifically comprises:

the first importance degree weight determining unit is used for determining the importance degree weight of the criterion layer by adopting an expert scoring method;

a second importance degree weight determination unit for determining an importance degree weight of the criterion layer with respect to the target layer based on the importance degree weight of the criterion layer;

a third importance degree weight determining unit, configured to determine an importance degree weight of the index layer relative to the criterion layer by using an expert scoring method;

a fourth importance weight determination unit for determining an importance weight of the index layer with respect to the target layer based on the importance weight of the criterion layer with respect to the target layer and the importance weight of the index layer with respect to the criterion layer;

a fifth importance level weight determining unit, configured to perform normalization processing on the importance level weight of the index layer with respect to the target layer to obtain an importance level weight of the hierarchical structure; the importance degree weight of the hierarchical structure is the importance degree weight of the risk elements relative to the target layer;

the first urgency weight determining unit is used for determining the urgency weight of the criterion layer by adopting an expert scoring method;

the second urgency degree weight determining unit is used for determining the urgency degree weight of the index layer relative to the criterion layer by adopting an expert scoring method based on the urgency degree weight of the criterion layer;

a third urgency weight determination unit for determining a urgency weight of the index layer relative to the target layer based on the urgency weight of the criterion layer and the urgency weight of the index layer relative to the criterion layer;

a fourth urgency weight determination unit, configured to perform normalization processing on the urgency weight of the index layer with respect to the target layer to obtain an urgency weight of the hierarchical structure; the urgency weight of the hierarchy is the urgency weight of the risk element relative to the target tier.

9. The risk management priority evaluation system for intelligent driving artificial intelligence of claim 6, wherein the management priority determination module specifically comprises:

a rank determination unit configured to determine an importance level of the risk element and an urgency level of the risk element according to the hierarchical importance weight and the hierarchical importance weight; the importance level comprises importance and next importance; the urgency class includes urgent and sub-urgent;

the important emergency matrix construction unit is used for constructing an important emergency matrix according to the importance degree grade of the risk element and the emergency degree grade of the risk element;

and the treatment priority determining unit is used for determining the treatment priority of the risk elements according to the important emergency matrix.

10. The risk management priority evaluation system for intelligent driving artificial intelligence according to claim 9, wherein the level determination unit specifically comprises:

the preset value acquiring subunit is used for acquiring a first preset value and a second preset value;

a first importance level determining subunit, configured to determine that the importance level of the risk element is important when the importance weight of the hierarchical structure is greater than a first preset value;

a second importance level determining subunit, configured to determine that the importance level of the risk element is the second importance when the importance level weight of the hierarchical structure is less than or equal to the first preset value;

a first urgency level determination subunit, configured to determine that the urgency level of the risk element is urgent when the urgency weight of the hierarchical structure is greater than a second preset value;

and a second urgency level determination subunit, configured to determine, when the urgency weight of the hierarchical structure is less than or equal to the second preset value, that the urgency level of the risk element is a second urgency.

Background

The intelligent driving is an important application field of artificial intelligence, and the current intelligent driving vehicle enters the stages of demonstration operation and product marketing, but the bottom layer data is heterogeneous, the algorithm is diversified, no clear standard is provided for governing, and the application has potential risks.

The world academy has a preliminary exploration on the theory of intelligent driving treatment, but the risk factors of the current intelligent driving are complex, and all countries have respective emphasis on the priority of treatment. At the federal level in the united states, the system provides regulatory principles for intelligent driving techniques; EU emphasis algorithm and data subject rights; germany focuses on the ethical aspects of intelligent driving. China also emphasizes the importance of building an artificial intelligence governance system in top-level strategies. However, at present, aiming at artificial intelligence risk management priority of intelligent driving, a targeted assessment method is not provided. Priority evaluation methods in other fields have strong industrial characteristics and barriers, and are difficult to directly transfer to the field of intelligent driving management; only by using an analytic hierarchy process, the evaluation by using the single dimension of the priority has strong subjectivity.

Disclosure of Invention

In order to solve the problems in the prior art, the invention provides a risk management priority evaluation method and system oriented to intelligent driving artificial intelligence.

In order to achieve the purpose, the invention provides the following scheme:

a risk management priority evaluation method facing intelligent driving artificial intelligence comprises the following steps:

acquiring technical requirements;

determining, based on the technical requirement, an outcome of not meeting the technical requirement;

analyzing and obtaining risk elements according to the consequences;

constructing a hierarchical structure for risk element treatment priority evaluation;

determining the emergency degree weight and the importance degree weight of the hierarchical structure by adopting an expert scoring method;

determining the treatment priority of the risk elements according to the emergency degree weight of the hierarchical structure and the importance degree weight of the hierarchical structure; the governance priorities include: a first priority, a second priority, and a third priority.

Preferably, the constructing a hierarchical structure of risk element governance priority assessment specifically includes:

constructing the hierarchical structure by taking the artificial intelligence risk management priority facing intelligent driving as a target layer, taking the risk category as a criterion layer and taking the risk elements as an index layer; the risk elements include algorithmic risk elements and data risk elements; the risk categories include algorithmic risk and data risk.

Preferably, the determining the urgency weight and the importance weight of the hierarchical structure by using an expert scoring method specifically includes:

determining the importance degree weight of the criterion layer by adopting an expert scoring method;

determining an importance weight of the criterion layer relative to the target layer based on the importance weight of the criterion layer;

determining the importance degree weight of the index layer relative to the criterion layer by adopting an expert scoring method;

determining a importance level weight of the index layer relative to the target layer based on the importance level weight of the criterion layer relative to the target layer and the importance level weight of the index layer relative to the criterion layer;

carrying out standardization processing on the importance degree weight of the index layer relative to the target layer to obtain the importance degree weight of the hierarchical structure; the importance degree weight of the hierarchical structure is the importance degree weight of the risk elements relative to the target layer;

determining the emergency degree weight of the criterion layer by adopting an expert scoring method;

determining the emergency degree weight of the index layer relative to the standard layer by adopting an expert scoring method based on the emergency degree weight of the standard layer;

determining an urgency weight of the index layer relative to the target layer based on the urgency weight of the criterion layer and the urgency weight of the index layer relative to the criterion layer;

carrying out standardization processing on the emergency degree weight of the index layer relative to the target layer to obtain the emergency degree weight of the hierarchical structure; the urgency weight of the hierarchy is the urgency weight of the risk element relative to the target tier.

Preferably, the determining the treatment priority of the risk element according to the emergency degree weight of the hierarchical structure and the importance degree weight of the hierarchical structure specifically includes:

determining an importance level of the risk element and an urgency level of the risk element according to the importance weight of the hierarchical structure and the importance weight of the hierarchical structure; the importance level comprises importance and next importance; the urgency class includes urgent and sub-urgent;

constructing an important emergency matrix according to the importance degree grade of the risk element and the emergency degree grade of the risk element;

and determining the treatment priority of the risk elements according to the important emergency matrix.

Preferably, the determining the importance level of the risk element and the urgency level of the risk element according to the importance weight of the hierarchical structure and the urgency weight of the hierarchical structure specifically includes:

acquiring a first preset value and a second preset value;

when the importance degree weight of the hierarchical structure is larger than a first preset value, the importance degree grade of the risk elements is important;

when the importance degree weight of the hierarchical structure is less than or equal to the first preset value, the importance degree grade of the risk elements is secondary importance;

when the emergency degree weight of the hierarchical structure is greater than a second preset value, the emergency degree grade of the risk elements is emergency;

and when the emergency degree weight of the hierarchical structure is less than or equal to the second preset value, the emergency degree grade of the risk elements is sub-emergency.

According to the specific embodiment provided by the invention, the invention discloses the following technical effects:

according to the risk management priority evaluation method facing the intelligent driving artificial intelligence, after the technical requirements are obtained, the consequences which do not meet the technical requirements are determined based on the technical requirements, then the risk elements are distinguished and analyzed according to the consequences, then after the hierarchical structure of the risk element management priority evaluation is established, the emergency degree weight and the importance degree weight of the hierarchical structure are determined by adopting an expert scoring method, finally, the management priority of the risk elements is determined according to the emergency degree weight of the hierarchical structure and the importance degree weight of the hierarchical structure, the subjectivity and the irrationality of single-dimension evaluation are overcome, clear and comprehensive evaluation of the risk elements is realized, and the evaluation method has the advantages of originality, scientificity, practicability and the like.

Corresponding to the risk management priority evaluation method facing the intelligent driving artificial intelligence, the invention also provides a specific implementation system, which comprises the following steps:

a risk management priority evaluation system oriented to intelligent driving artificial intelligence comprises:

the technical requirement acquisition module is used for acquiring technical requirements;

an outcome determination module to determine an outcome that does not meet the specification based on the specification;

the risk element analysis module is used for analyzing and obtaining a risk element according to the result;

the hierarchical structure construction module is used for constructing a hierarchical structure for risk element treatment priority evaluation;

the weight determining module is used for determining the emergency degree weight and the importance degree weight of the hierarchical structure by adopting an expert scoring method;

the treatment priority determining module is used for determining the treatment priority of the risk elements according to the emergency degree weight of the hierarchical structure and the importance degree weight of the hierarchical structure; the governance priorities include: a first priority, a second priority, and a third priority.

Preferably, the hierarchical structure building module specifically includes:

the hierarchical structure building unit is used for building the hierarchical structure by taking the artificial intelligence risk management priority facing intelligent driving as a target layer, taking the risk category as a criterion layer and taking the risk elements as an index layer; the risk elements include algorithmic risk elements and data risk elements; the risk categories include algorithmic risk and data risk.

Preferably, the weight determining module specifically includes:

the first importance degree weight determining unit is used for determining the importance degree weight of the criterion layer by adopting an expert scoring method;

a second importance degree weight determination unit for determining an importance degree weight of the criterion layer with respect to the target layer based on the importance degree weight of the criterion layer;

a third importance degree weight determining unit, configured to determine an importance degree weight of the index layer relative to the criterion layer by using an expert scoring method;

a fourth importance weight determination unit for determining an importance weight of the index layer with respect to the target layer based on the importance weight of the criterion layer with respect to the target layer and the importance weight of the index layer with respect to the criterion layer;

a fifth importance level weight determining unit, configured to perform normalization processing on the importance level weight of the index layer with respect to the target layer to obtain an importance level weight of the hierarchical structure; the importance degree weight of the hierarchical structure is the importance degree weight of the risk elements relative to the target layer;

the first urgency weight determining unit is used for determining the urgency weight of the criterion layer by adopting an expert scoring method;

the second urgency degree weight determining unit is used for determining the urgency degree weight of the index layer relative to the criterion layer by adopting an expert scoring method based on the urgency degree weight of the criterion layer;

a third urgency weight determination unit for determining a urgency weight of the index layer relative to the target layer based on the urgency weight of the criterion layer and the urgency weight of the index layer relative to the criterion layer;

a fourth urgency weight determination unit, configured to perform normalization processing on the urgency weight of the index layer with respect to the target layer to obtain an urgency weight of the hierarchical structure; the urgency weight of the hierarchy is the urgency weight of the risk element relative to the target tier.

Preferably, the governing priority determining module specifically includes:

a rank determination unit configured to determine an importance level of the risk element and an urgency level of the risk element according to the hierarchical importance weight and the hierarchical importance weight; the importance level comprises importance and next importance; the urgency class includes urgent and sub-urgent;

the important emergency matrix construction unit is used for constructing an important emergency matrix according to the importance degree grade of the risk element and the emergency degree grade of the risk element;

and the treatment priority determining unit is used for determining the treatment priority of the risk elements according to the important emergency matrix.

Preferably, the level determining unit specifically includes:

the preset value acquiring subunit is used for acquiring a first preset value and a second preset value;

a first importance level determining subunit, configured to determine that the importance level of the risk element is important when the importance weight of the hierarchical structure is greater than a first preset value;

a second importance level determining subunit, configured to determine that the importance level of the risk element is the second importance when the importance level weight of the hierarchical structure is less than or equal to the first preset value;

a first urgency level determination subunit, configured to determine that the urgency level of the risk element is urgent when the urgency weight of the hierarchical structure is greater than a second preset value;

and a second urgency level determination subunit, configured to determine, when the urgency weight of the hierarchical structure is less than or equal to the second preset value, that the urgency level of the risk element is a second urgency.

Because the technical effect achieved by the risk management priority evaluation system facing the intelligent driving artificial intelligence provided by the invention is the same as the technical effect achieved by the risk management priority evaluation method facing the intelligent driving artificial intelligence provided by the invention, the details are not repeated herein.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.

FIG. 1 is a flow chart of a risk management priority evaluation method for intelligent driving artificial intelligence provided by the invention;

FIG. 2 is a frame diagram of an implementation of the risk management priority evaluation method for intelligent driving artificial intelligence provided by the invention;

FIG. 3 is a hierarchical diagram of a risk element priority assessment provided by the practice of the present invention;

fig. 4 is an important emergency matrix diagram of the priority of intelligent driving artificial intelligence risk management provided by the embodiment of the present invention;

fig. 5 is a schematic structural diagram of the risk management priority evaluation system for intelligent driving artificial intelligence provided by the invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

The invention aims to solve the problem that risks and treatment priorities are not clear in the process of establishing an intelligent driving artificial intelligent treatment framework in the prior art, and establishes a risk treatment priority evaluation method and a risk treatment priority evaluation system which have the advantages of originality, scientificity, practicability and the like and are oriented to intelligent driving artificial intelligence.

In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.

Based on the implementation architecture shown in fig. 2, the invention provides a risk management priority evaluation method facing intelligent driving artificial intelligence, and as shown in fig. 1, the risk management priority evaluation method includes:

step 100: and acquiring technical requirements. Specifically, based on a constructed risk library facing intelligent driving artificial intelligence management, the technical requirements of an intelligent driving artificial intelligence algorithm and data are combed. The specification classification is shown in table 3 below.

Step 101: the consequence of not meeting the specification is determined based on the specification. Specifically, according to the technical requirements, the potential consequences caused by the technical requirements not being met are combed. For example, in the technical requirements of a positioning algorithm, the positioning accuracy of the positioning algorithm needs to be guaranteed to a certain level, otherwise, the obstacle avoidance capability of the vehicle is obstructed, and thus, the result of an accident is caused.

Step 102: and (5) obtaining risk factors according to result identification. Based on the example in step 101, an "algorithm accuracy risk" can be identified. Further to assess the stringency, a risk factor may be defined, for example, an algorithm accuracy risk, which may be defined as "risk of large deviation between the output value of the intelligent driving algorithm and the true or optimal value".

Step 103: and constructing a hierarchical structure of risk element governance priority evaluation. Specifically, a hierarchical structure is constructed by taking the artificial intelligence risk management priority facing intelligent driving as a target layer, taking risk categories as a criterion layer and taking risk elements as an index layer. The risk elements include algorithmic risk elements and data risk elements. The risk categories include algorithmic risk and data risk.

Step 104: and determining the emergency degree weight and the importance degree weight of the hierarchical structure by adopting an expert scoring method.

Step 105: and determining the treatment priority of the risk elements according to the emergency degree weight of the hierarchical structure and the importance degree weight of the hierarchical structure. Governing the priority includes: a first priority, a second priority, and a third priority.

Further, the specific implementation process of the step 104 is as follows:

step 1041: and determining the importance degree weight of the criterion layer by adopting an expert scoring method. Specifically, the method comprises the following steps:

and (3) constructing a judgment matrix, establishing a paired comparison matrix by adopting a 1-9 scale method aiming at the importance degree of algorithm and data risk management in a pairwise comparison mode, and performing pairwise importance comparison by expert scoring.

The 1-9 scale method used for the comparison of importance is shown in Table 1, wherein b_ijIndicating the importance of i to j, the scale values may be represented by intermediate numbers for the cases between the determinations in table 1.

TABLE 1 Risk index importance Table

Comparison of importance	bij
		i and j are of equal importance	1
i is slightly more important than j	3
		i is significantly more important than j	5
i is strongly important than j	7
		i is extremely important than j	9

All comparison results can be used as comparison matrix B_iRepresents:

wherein, b_ijRepresenting the risk importance of the algorithm risk i relative to the data risk j in the criterion layer. When j is compared with i, it can be represented by the reciprocal of the comparison scale value of i and j. When i is equal to j, b_ij＝1。

Step 1042: the importance level weight of the criterion layer relative to the target layer is determined based on the importance level weight of the criterion layer. Specifically, the method comprises the following steps:

for comparison matrix B_iCalculating the importance degree weight of the criterion layer relative to the target layer as follows:

in the formula: b_iiIs the weight of the criterion layer relative to the target layer. b_ijIn the layer of presentation criteriaRisk importance of algorithm risk i relative to data risk j, b_ijAlso elements in decision matrix a.

Step 1043: and determining the importance degree weight of the index layer relative to the standard layer by adopting an expert scoring method.

Specifically, the method comprises the following steps:

and constructing a judgment matrix, establishing paired comparison matrixes by adopting a 1-9 scale method aiming at the importance degree of treatment of a plurality of risk elements on respective index layers of data and algorithm risks in a pairwise comparison mode, carrying out pairwise importance comparison through expert scoring, and carrying out consistency inspection on the judgment matrixes.

All comparison results thereof can be used as a comparison matrix C_iRepresents:

wherein c is_ijRepresenting the risk importance of i relative to j, j when compared to i may be represented by the reciprocal of the scale value of i compared to j, and n represents the total number of risk elements at a particular criteria level (data or algorithm).

For comparison matrix C_iCalculating the weight of each risk element relative to the criterion layer:

in the formula: c. C_iiThe importance of each risk element relative to the criteria layer is weighted.

Step 1044: the importance degree weight of the index layer relative to the target layer is determined based on the importance degree weight of the criterion layer relative to the target layer and the importance degree weight of the index layer relative to the criterion layer. Specifically, the method comprises the following steps:

and (5) performing consistency check on the judgment matrix in the step 1043.

And (3) carrying out consistency check on the judgment matrix, and calculating a consistency index CI (consistency index) formula as follows:

wherein λ is_maxThe maximum eigenvalue of the matrix a is judged. Then, the consistency ratio cr (consistency ratio) is calculated:

where RI is a random consistency index and can be obtained by looking up table 2.

TABLE 2 RI LUT

And when CR is less than or equal to 0.10, the consistency of the judgment matrix is considered to be acceptable, otherwise, expert scoring is carried out again, and the judgment matrix is properly corrected until the consistency passes the consistency check.

The importance degree weight of the index layer relative to the target layer is: w is a_ii＝b_ii×c_ii

In the formula: w is a_iiThe importance level of each risk element relative to the target layer is weighted. b_iiThe importance degree of the two risk categories of the criterion layer algorithm and the data relative to the target layer is weighted. c. C_iiThe importance of each risk element relative to the criteria layer is weighted.

Step 1045: and carrying out standardization processing on the importance degree weight of the index layer relative to the target layer to obtain the importance degree weight of the hierarchical structure. The importance degree weight of the hierarchical structure is the importance degree weight of the risk elements relative to the target layer. Specifically, the method comprises the following steps:

the formula for normalizing the importance degree weight is as follows:

in the formula: w is a_niiThe importance degree weight of each risk element relative to the target layer after standardization. μ is the average of the importance weights of the individual risk elements relative to the target layer. σ is the standard deviation of the importance weighting of each risk element relative to the target layer.

Step 1046: and determining the emergency degree weight of the criterion layer by adopting an expert scoring method. Specifically, a judgment matrix is constructed, a pairwise comparison matrix is established by a 1-9 scale method aiming at the emergency degree of algorithm and data risk management in a pairwise comparison mode, pairwise emergency degree comparison is carried out through expert scoring, and the method is the same as the step 1041.

Step 1047: and determining the emergency degree weight of the index layer relative to the standard layer by adopting an expert scoring method based on the emergency degree weight of the standard layer. Specifically, the urgency degree weight of the index layer relative to the criterion layer is determined. And (3) constructing a judgment matrix, establishing a paired comparison matrix by adopting a 1-9 scale method aiming at the emergency degree of treatment of a plurality of risk elements on respective index layers of data and algorithm risks in a pairwise comparison mode, and performing pairwise emergency degree comparison by expert scoring, wherein the method is the same as the step 1042.

Step 1048: and determining the urgency weight of the index layer relative to the target layer based on the urgency weight of the criterion layer and the urgency weight of the index layer relative to the criterion layer. Specifically, the consistency of the judgment matrix in step 1047 is checked, and the method is the same as that in step 1043. Determining the urgency weight of the index layer relative to the target layer: w is a_ei＝b_ei×c_ei。

In the formula, w_eiFor the urgency weight of each risk element relative to the target layer, b_eiEmergency degree weight of two risk categories of the criterion layer algorithm and data relative to the target layer, c_eiThe urgency of each risk element is weighted against the criteria level.

Step 1049: and carrying out standardization processing on the emergency degree weight of the index layer relative to the target layer to obtain the emergency degree weight of the hierarchical structure. The urgency weight of the hierarchy is the urgency weight of the risk element relative to the target tier. Specifically, the method comprises the following steps:

and (3) carrying out standardization processing on the emergency degree weight:

in the formula: w is a_neiIs the urgency weight of each risk element relative to the target tier after normalization. μ is the average of the urgency weights of the individual risk elements relative to the target layer. σ is the standard deviation of the urgency weights of the individual risk elements relative to the target tier.

Based on the importance degree weight of the hierarchical structure determined in step 104, the specific implementation process of determining the governance priority of the risk elements in step 105 is as follows:

step 1051: and determining the importance degree grade of the risk element and the urgency degree grade of the risk element according to the urgency degree weight of the hierarchical structure and the importance degree weight of the hierarchical structure. The importance level includes importance and sub-importance. The level of urgency includes urgent and sub-urgent. Specifically, the method comprises the following steps:

and acquiring a first preset value and a second preset value. The first preset value and the second preset value are both 0 in the present invention.

And when the importance degree weight of the hierarchical structure is greater than a first preset value, the importance degree grade of the risk elements is important.

And when the importance degree weight of the hierarchical structure is less than or equal to a first preset value, the importance degree grade of the risk elements is the secondary importance.

And when the emergency degree weight of the hierarchical structure is greater than a second preset value, the emergency degree grade of the risk elements is emergency.

And when the emergency degree weight of the hierarchical structure is less than or equal to a second preset value, the emergency degree grade of the risk elements is sub-emergency.

Step 1052: and constructing an important emergency matrix according to the importance degree grades of the risk elements and the emergency degree grades of the risk elements.

Step 1053: and determining the treatment priority of the risk elements according to the important emergency matrix. Wherein important urgency is a first priority, important suburgency and urgent suburgency are a second priority, and subimportance suburgency is a third priority.

The following describes a specific implementation process of the risk management priority evaluation method for intelligent driving artificial intelligence provided by the invention.

The method comprises the following steps: the method comprises the following steps of constructing a risk library for intelligent driving artificial intelligent management and establishing a hierarchical structure for risk element management priority evaluation, wherein the hierarchical structure comprises the following steps:

11) and combing the technical requirements of the artificial intelligent data of the intelligent driving, and distinguishing and analyzing potential risk factors caused by the fact that the technical requirements are not met according to the technical requirements of the data. As shown in table 3.

TABLE 3 risk analysis table for intelligent driving artificial intelligence data

12) And combing the technical requirements of the artificial intelligence algorithm of the intelligent driving, and distinguishing and analyzing potential risk factors caused by the fact that the technical requirements are not met according to the technical requirements of the algorithm. As shown in table 4.

Table 4 risk analysis table of intelligent driving artificial intelligence algorithm

13) The identified risk elements are defined as shown in table 5.

TABLE 5 Intelligent Driving Artificial Intelligence Risk elements definition Table

Step two: hierarchical structure for constructing risk element governance priority assessment

21) The target layer is an evaluation target, namely a risk management priority facing intelligent driving artificial intelligence, the criterion layer is a data risk and an algorithm risk, the index layer is various risk elements defined in the step 13), and the constructed hierarchical structure is shown in fig. 3.

211) Determining the importance degree weight of the criterion layer, constructing a judgment matrix by an expert scoring method, and comparing the importance degree by adopting a 1-9 scale method, wherein the meanings of the importance degree are shown in table 1. All the comparison results can be used in the comparison matrix B_iAnd (4) showing. Judgment matrix B_iAs shown in table 6.

TABLE 6 weight judgment matrix table for importance of criterion layer

	Data of	Algorithm
			Data of	1.00	0.67
Algorithm	1.50	1.00

Calculating to obtain the importance degree weight b of the data risk relative to the target layer_i10.40, importance weighting b of algorithmic risk with respect to target layer_i2＝0.60。

212) Determining importance of an index layer relative to a criteria layerDegree weight. And establishing a pair of judgment matrixes by adopting a 1-9 scale method, wherein the meanings of the judgment matrixes are shown in the table 1. The significance of each two is compared by expert scoring, and all comparison results can be used by the judgment matrix C_iAnd (4) showing. The numbering of the risk elements is detailed in the hierarchy of risk management priority assessment as shown in fig. 3. Data risk is shown in table 7 and algorithm risk is shown in table 8.

Table 7 data risk indicator layer importance degree weight judgment matrix table

TABLE 8 Algorithm Risk indicator layer importance weight judgment matrix Table

For comparison matrix C_iUsing the above calculation C_iiThe formula (c) calculates the weight of each risk element relative to the criteria layer. The importance weights of the risks with respect to the criteria layer are calculated as shown in table 9.

TABLE 9 importance weighting Table for risks versus level of criteria

And carrying out consistency check on the judgment matrix in the step 212).

The data risk index layer importance degree weight judgment matrix consistency test result is as follows:

CR is less than or equal to 0.10, which indicates that the consistency of the judgment matrix is acceptable.

The test result of the consistency of the importance degree weight judgment matrix of the algorithm risk index layer is as follows:

CR is less than or equal to 0.10, which indicates that the consistency of the judgment matrix is acceptable.

214) Determining importance weighting of the index layer relative to the target layer.

215) The importance degree weights were normalized, and the calculation results are shown in table 10.

Table 10 importance degree weight calculation table of index layer with respect to target layer

22) Calculating an urgency weight for each risk element

221) And determining the emergency degree weight of the criterion layer, constructing a judgment matrix by an expert scoring method, and comparing the emergency degree by adopting a 1-9 scale method, wherein the meanings of the emergency degree weight are shown in table 1. All comparison results can be used as comparison matrix B_eAnd (4) showing. Judgment matrix B_eAs shown in table 11:

TABLE 11 criterion layer urgency weight determination matrix table

	Data of	Algorithm
			Data of	1.00	0.77
Algorithm	1.30	1.00

Calculating to obtain the urgency weight b of the data risk relative to the target layer_e10.435, the urgency weight b of the algorithmic risk with respect to the target layer_e2＝0.565。

222) An urgency weight of the index layer relative to the criterion layer is determined. And establishing a pair of judgment matrixes by adopting a 1-9 scale method, wherein the meanings of the judgment matrixes are shown in the table 1. Every two urgency comparisons are carried out by expert scoring, and all comparison results can be used as a judgment matrix C_eAnd (4) showing. Wherein the data risk is shown in table 12 and the algorithm risk is shown in table 13.

Table 12 data risk indicator layer emergency degree weight judgment matrix table

Table 13 algorithm risk indicator layer emergency weight judgment matrix table

For decision matrix C_eThe emergency weight of each risk element with respect to the criterion layer is calculated, and the emergency weight of each risk with respect to the criterion layer is calculated as shown in table 14.

TABLE 14 urgency weighting tables for risks versus criteria level

213) And carrying out consistency check on the judgment matrix in the step 222).

The consistency test result of the data risk index layer emergency degree weight judgment matrix is as follows:

CR is less than or equal to 0.10, which indicates that the consistency of the judgment matrix is acceptable.

The consistency test result of the emergency degree weight judgment matrix of the algorithm risk index layer is as follows:

CR is less than or equal to 0.10, which indicates that the consistency of the judgment matrix is acceptable.

214) Determining an urgency weight w for an index layer relative to a target layer_ei。

215) The emergency weight was normalized, and the calculation results are shown in table 15.

Table 15 emergency degree weight calculation table of index layer relative to target layer

Elements of risk	bei	cei	wei	wnei
					Data security risk	0.435	0.35	0.15	-0.67
Risk of data integrity	0.435	0.09	0.04	0.71
					Data standardRisk of certainty	0.435	0.19	0.08	-0.77
Data timeliness risk	0.435	0.08	0.03	-1.41
					Risk of data ownership	0.435	0.03	0.01	0.28
Data consistency risk	0.435	0.16	0.07	-0.47
					Data privacy risk	0.435	0.10	0.04	1.59
Risk of algorithm precision	0.565	0.26	0.15	-0.15
					AlgorithmRisk of robustness	0.565	0.13	0.07	-1.03
Algorithmic ethical risk	0.565	0.06	0.03	0.50
					Risk of algorithm efficiency	0.565	0.18	0.10	-0.58
Algorithmic privacy risk	0.565	0.09	0.05	1.99
					Algorithmic anthropomorphic risk	0.565	0.29	0.16	-0.67

Step three: constructing an important emergency matrix, and sequencing the treatment priority of each risk element, wherein the method comprises the following steps:

31) and determining the importance degree/emergency degree grade according to the standardized importance degree/emergency degree weight of each risk element, wherein the importance degree weight is more than 0 and less than 0. The urgency degree weight is more than 0, and the second urgency is less than 0.

32) And constructing an important emergency matrix according to the importance/emergency degree grades of the risk elements.

The critical emergency matrix created according to the above steps is shown in fig. 4.

33) Judging the treatment priority according to the important emergency matrix: where important emergency is a first priority, important sub-emergency and urgent sub-importance are a second priority, and sub-important sub-emergency is a third priority, as shown in table 16.

TABLE 16 Risk management priority evaluation result table for intelligent driving artificial intelligence

In summary, compared with the prior art, the invention also has the following advantages:

1. originality: at present, the establishment of an artificial intelligent management framework facing intelligent driving is urgently needed in China, and an evaluation method of corresponding risk management priority is used as an important premise for establishing the framework, and the establishment of the framework still needs to be deepened at present. The method is developed from two dimensions of importance and urgency based on an analytic hierarchy process and combined with expert knowledge, a set of risk factor distinguishing and evaluating method facing intelligent driving artificial intelligent management is creatively established, and research and construction of an intelligent driving artificial intelligent management framework in China can be effectively enabled.

2. Scientifically: the invention provides an artificial intelligence risk management priority evaluation method for intelligent driving, which is used for identifying risk elements and constructing a risk library based on the technical requirements of an intelligent driving algorithm and data. The priority weights are evaluated from two aspects of importance and urgency, so that an important emergency matrix is constructed, the treatment priority is finally determined, clear and comprehensive risk elements can be realized, and the priority evaluation is scientific and reasonable.

3. The practicability is as follows: the risk treatment priority evaluation method facing the intelligent driving artificial intelligence is a set of systematic flow, can efficiently evaluate the treatment priority of risk elements, and has wide applicability in the aspect of facing the intelligent driving artificial intelligence treatment.

In addition, corresponding to the above-mentioned risk management priority evaluation method for intelligent driving artificial intelligence, the present invention also provides a risk management priority evaluation system for intelligent driving artificial intelligence, as shown in fig. 5, the risk management priority evaluation system includes: the system comprises a technical requirement acquisition module 1, an outcome determination module 2, a risk element distinguishing module 3, a hierarchical structure construction module 4, a weight determination module 5 and a treatment priority determination module 6.

The technical requirement obtaining module 1 is used for obtaining the technical requirement.

The consequence determination module 2 is used to determine the consequence of not meeting the specification based on the specification.

The risk element analyzing module 3 is used for analyzing and obtaining the risk elements according to the results.

The hierarchical structure building module 4 is used for building a hierarchical structure of risk element governance priority evaluation.

The weight determination module 5 is used for determining the urgency weight and the importance weight of the hierarchical structure by adopting an expert scoring method.

And the treatment priority determining module 6 is used for determining the treatment priority of the risk elements according to the emergency degree weight of the hierarchical structure and the importance degree weight of the hierarchical structure. Governing the priority includes: a first priority, a second priority, and a third priority.

Further, the above-mentioned adopted hierarchical structure building module 4 preferably includes: and a hierarchical structure building unit.

The hierarchical structure building unit is used for building a hierarchical structure by taking the artificial intelligence risk management priority facing intelligent driving as a target layer, taking the risk category as a criterion layer and taking the risk elements as an index layer. The risk elements include algorithmic risk elements and data risk elements. The risk categories include algorithmic risk and data risk.

Further, the weight determining module 5 preferably includes: the first importance degree weight determining unit, the second importance degree weight determining unit, the third importance degree weight determining unit, the fourth importance degree weight determining unit, the fifth importance degree weight determining unit, the first urgency degree weight determining unit, the second urgency degree weight determining unit, the third urgency degree weight determining unit and the fourth urgency degree weight determining unit.

The first importance degree weight determining unit is used for determining importance degree weights of the criterion layers by adopting an expert scoring method.

The second importance level weight determination unit is configured to determine an importance level weight of the criterion layer with respect to the target layer based on the importance level weight of the criterion layer.

The third importance degree weight determining unit is used for determining the importance degree weight of the index layer relative to the criterion layer by adopting an expert scoring method.

The fourth importance weight determination unit is configured to determine an importance weight of the index layer with respect to the target layer based on the importance weight of the criterion layer with respect to the target layer and the importance weight of the index layer with respect to the criterion layer.

The fifth importance level weight determining unit is used for carrying out standardization processing on the importance level weight of the index layer relative to the target layer to obtain the importance level weight of the hierarchical structure. The importance degree weight of the hierarchical structure is the importance degree weight of the risk elements relative to the target layer.

The first urgency weight determination unit is used for determining the urgency weight of the criterion layer by adopting an expert scoring method.

The second urgency weight determination unit is used for determining the urgency weight of the index layer relative to the criterion layer by adopting an expert scoring method based on the urgency weight of the criterion layer.

The third urgency weight determination unit is used for determining the urgency weight of the index layer relative to the target layer based on the urgency weight of the criterion layer and the urgency weight of the index layer relative to the criterion layer.

The fourth urgency weight determination unit is used for carrying out standardization processing on the urgency weight of the index layer relative to the target layer to obtain the urgency weight of the hierarchical structure. The urgency weight of the hierarchy is the urgency weight of the risk element relative to the target tier.

Further, the adopted treatment priority determining module 6 specifically includes: the system comprises a grade determining unit, an important emergency matrix constructing unit and a governing priority determining unit.

The level determining unit is used for determining the importance level of the risk element and the urgency level of the risk element according to the urgency weight of the hierarchical structure and the importance weight of the hierarchical structure. The importance level includes importance and sub-importance. The level of urgency includes urgent and sub-urgent.

The important emergency matrix building unit is used for building an important emergency matrix according to the importance degree grades of the risk elements and the emergency degree grades of the risk elements.

And the treatment priority determining unit is used for determining the treatment priority of the risk elements according to the important emergency matrix.

Wherein, the grade determining unit preferably includes: the emergency degree determination device comprises a preset value acquisition subunit, a first importance degree level determination subunit, a second importance degree level determination subunit, a first emergency degree level determination subunit and a second emergency degree level determination subunit.

The preset value acquiring subunit is used for acquiring a first preset value and a second preset value.

The first importance level determines that the importance level of the risk elements is important when the importance weight of the subunit for the hierarchical structure is greater than a first preset value.

The second importance level determines that the importance level of the risk elements is the second importance when the importance weight of the subunit for the hierarchical structure is less than or equal to the first preset value.

The first urgency level determines that the urgency level of the risk element is urgent when the urgency weight of the subunit for the hierarchical structure is greater than a second preset value.

And when the second urgency level determination subunit is used for determining that the urgency weight of the hierarchical structure is less than or equal to a second preset value, the urgency level of the risk element is a second urgency.

The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description.

The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.