Enterprise violation risk prediction model training method and related device

Abstract

The invention provides an enterprise violation risk prediction model training method and a related device, and the method comprises the steps: obtaining enterprise contract data, inputting the enterprise contract data into a pre-trained enterprise contract data expansion model, and obtaining expanded contract data; combining the enterprise contract data and the expanded contract data to obtain a first contract sample set; performing feature extraction on the first contract sample set by adopting a pre-trained feature extraction model to obtain contract features of each contract sample in the first contract sample set; according to the risk type of the contract sample, correspondingly marking contract features, and combining all marked contract features to obtain a training feature set; training the classification model by adopting the training feature set to obtain an enterprise violation risk prediction model; according to the method, the enterprise violation risk prediction model for performing high-accuracy judgment on the enterprise contract risk can be obtained through training.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of data processing technology, and in particular to a training method and related device for an enterprise violation risk prediction model. Background Art

[0002] As corporate compliance requirements become increasingly stringent, contract management has gradually become an important part of corporate operations. During the contract signing process, companies need to comprehensively evaluate the compliance of contract terms to reduce potential legal risks and operating losses. Traditional contract compliance audits rely on manual operations, which are time-consuming and easily affected by subjective factors, and are difficult to cover the risk clauses and potential violations hidden in the contract.

[0003] In the existing technology, the detection method based on deep learning is restricted by the scale and quality of training data, and it is difficult to fully explore the complex semantics and logical associations in the contract terms. The prediction accuracy of the trained enterprise violation risk prediction model is low. Summary of the invention

[0004] The embodiments of the present invention provide a training method and related devices for an enterprise violation risk prediction model to solve the problem that the existing deep learning-based detection method is restricted by the scale and quality of training data, making it difficult to fully explore the complex semantics and logical associations in the contract terms, and the trained enterprise violation risk prediction model has low prediction accuracy.

[0005] In order to solve the above-mentioned technical problems, the present invention is achieved as follows: In a first aspect, an embodiment of the present invention provides a method for training an enterprise violation risk prediction model, comprising: Acquire enterprise contract data, and input the enterprise contract data into a pre-trained enterprise contract data expansion model to obtain expanded contract data; Combining the enterprise contract data with the extended contract data to obtain a first contract sample set; Using a pre-trained feature extraction model to extract features from the first contract sample set, to obtain contract features of each contract sample in the first contract sample set; labeling the contract features according to the risk type of the contract sample, and combining all the labeled contract features to obtain a training feature set; The classification model is trained using the training feature set to obtain an enterprise violation risk prediction model; Among them: the enterprise contract data expansion model is obtained by training based on an adversarial network algorithm model generated based on a dual discriminator; the feature extraction model is obtained by training based on a neural network model optimized by multi-dimensional sonic bats.

[0006] Optionally, the training process of the enterprise contract data expansion model includes: After completing the first round of training iterations for the adversarial network algorithm model, in each round of training iterations, the noise vector of the current round of training iterations is corrected according to the contract data generated by the adversarial network algorithm model that completed the previous round of training iterations.

[0007] Optionally, in the adversarial network algorithm model, one discriminator is used to judge the authenticity of the contract data generated by the adversarial network algorithm model, and another discriminator is used to judge the category distribution of the contract data generated by the adversarial network algorithm model.

[0008] Optionally, each time a training iteration is completed for the adversarial network algorithm model, it is verified whether the total number of current training iterations exceeds a preset iteration number threshold, and a verification result is obtained; If the verification result indicates that the total number of current training iterations is less than the iteration number threshold, continue to perform the next round of training iterations on the adversarial network algorithm model; If the verification result indicates that the total number of current training iterations is greater than or equal to the iteration number threshold, the next round of iterative training is stopped, and the adversarial network algorithm model after this training iteration is used as the enterprise contract data expansion model.

[0009] Optionally, the training method of the feature extraction model includes: After each training iteration of the neural network model, determine whether the neural network model after the training iteration meets the convergence condition, and obtain a determination result; If the judgment result indicates that the convergence condition is not met, continue to perform the next round of training iteration on the neural network model; If the judgment result indicates that the convergence condition is met, the next round of iterative training is stopped, and the neural network model after the current training iteration is used as the feature extraction model; The expression of the convergence condition is as follows:

[0010] In the formula, is the maximum value of all parameter and bias updates in the neural network model; is the convergence threshold hyperparameter.

[0011] Optionally, the training method of the feature extraction model includes: In the multi-dimensional sound wave modulation link, the sound wave attribute of each parameter in the neural network is corrected according to the performance of the neural network after the most recent training iteration at the current moment; The expression of the correction process is as follows:

[0012] In the formula, is the parameter of the neural network in the The frequency at the iteration; is the parameter of the neural network in the The frequency at the iteration; is the partial derivative of the loss function of the neural network with respect to the weight parameter; is the loss function of the neural network; is the frequency regulation rate; is the parameter of the neural network in the The frequency update increment for each iteration.

[0013] In a second aspect, an embodiment of the present invention provides a training device for an enterprise violation risk prediction model, comprising: An expansion module, used for acquiring enterprise contract data, and inputting the enterprise contract data into a pre-trained enterprise contract data expansion model to obtain expanded contract data; A combination module, used for combining the enterprise contract data and the extended contract data to obtain a first contract sample set; a feature extraction module, configured to extract features from the first contract sample set using a pre-trained feature extraction model to obtain contract features of each contract sample in the first contract sample set; annotate the contract features according to the risk type of the contract sample, and combine all the annotated contract features to obtain a training feature set; A training module is used to train the classification model using a training feature set to obtain an enterprise violation risk prediction model; Among them: the enterprise contract data expansion model is obtained by training based on an adversarial network algorithm model generated based on a dual discriminator; the feature extraction model is obtained by training based on a neural network model optimized by multi-dimensional sonic bats.

[0014] In a third aspect, an embodiment of the present invention provides an electronic device comprising a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction, when executed by the processor, implements the steps in the method for training an enterprise violation risk prediction model as described in any one of the first aspects.

[0015] In a fourth aspect, an embodiment of the present invention provides a readable storage medium, on which a program or instruction is stored, and when the program or instruction is executed by a processor, the steps in the training method of the enterprise violation risk prediction model as described in any one of the first aspects are implemented.

[0016] In a fifth aspect, an embodiment of the present invention provides a computer program product, comprising computer instructions, which, when executed by a processor, implement the steps of the method for training an enterprise violation risk prediction model as described in any one of the first aspects.

[0017] In an embodiment of the present invention, the enterprise contract data is obtained and input into a pre-trained enterprise contract data expansion model to obtain the expanded contract data; the enterprise contract data and the expanded contract data are combined to obtain a first contract sample set; the pre-trained feature extraction model is used to extract features from the first contract sample set to obtain the contract features of each contract sample in the first contract sample set; the contract features are labeled according to the risk type of the contract sample, and all the labeled contract features are combined to obtain a training feature set; the classification model is trained using the training feature set to obtain an enterprise violation risk prediction model; wherein: the enterprise contract data expansion model is trained based on an adversarial network algorithm model generated based on a dual discriminator; the feature extraction model is trained based on a neural network model optimized by a multi-dimensional sonic bat It is obtained that the enterprise contract data expansion model obtained by training according to the adversarial network algorithm model generated based on the dual discriminator realizes the expansion of the enterprise contract data and increases the scale of training data; and, the feature extraction model obtained by training according to the neural network model optimized by multi-dimensional sonic bats is used to extract features of the first contract sample set, and the training feature set is obtained by annotating based on the extracted contract features, thereby improving the quality of the training data (i.e., the training feature set) used to train the classification model, and then improving the classification model's cognitive accuracy of the sample features. The trained enterprise violation risk prediction model can fully explore the complex semantics and logical associations in the contract terms, and then obtain accurate violation risk prediction results. The present invention can train an enterprise violation risk prediction model that can make high-accuracy judgments on enterprise contract risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings: Figure 1 A flowchart of a method for training an enterprise violation risk prediction model according to an embodiment of the present invention; Figure 2 A schematic diagram of the principle of a method for predicting and warning corporate compliance violations based on deep learning; Figure 3 The principle block diagram of the training device for the enterprise violation risk prediction model; Figure 4The figure is a principle block diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The following will be combined with the accompanying drawings in the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] The terms "first", "second", etc. in the embodiments of the present invention are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the terms used in this way are interchangeable where appropriate, so that the embodiments of the present invention can be implemented in an order other than those illustrated or described herein, and the objects distinguished by "first" and "second" are generally of the same type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "or" in the embodiments of the present invention represents at least one of the connected objects. For example, "A or B" covers three schemes, namely, Scheme 1: including A but not including B; Scheme 2: including B but not including A; Scheme 3: including both A and B. The character " / " generally indicates that the objects associated with each other are in an "or" relationship.

[0021] In the technical solutions of the embodiments of the present invention, words such as “connect”, “couple” or “connected” are not limited to physical or mechanical connections, but may include electrical connections.

[0022] In addition, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0023] It should be noted that the collection, collection, updating, analysis, processing, use, transmission, storage and other aspects of personal information involved in the technical solutions of the embodiments of the present invention are in compliance with the provisions of relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. Necessary measures are taken for personal information to prevent illegal access to personal information data and maintain personal information security and network security.

[0024] The present invention provides a method for training an enterprise violation risk prediction model. Figure 1 As shown, Figure 1 The flowchart of the training method of the enterprise violation risk prediction model according to an embodiment of the present invention includes: Step 11: Obtain enterprise contract data, and input the enterprise contract data into a pre-trained enterprise contract data expansion model to obtain expanded contract data; Step 12: Combine the enterprise contract data and the expanded contract data to obtain a first contract sample set; Step 13: Use the pre-trained feature extraction model to extract features from the first contract sample set to obtain contract features of each contract sample in the first contract sample set; annotate the contract features according to the risk type of the contract sample, and combine all the annotated contract features to obtain a training feature set; Step 14: Use the training feature set to train the classification model to obtain an enterprise violation risk prediction model; Among them: the enterprise contract data expansion model is trained based on the adversarial network algorithm model generated by the dual discriminator; the feature extraction model is trained based on the neural network model optimized by multi-dimensional sonic bats.

[0025] In the embodiment of the present invention, the enterprise contract data is real contract data that is collected, and the enterprise contract data can be obtained from a storage medium that stores the enterprise contract data.

[0026] In the embodiment of the present invention, the contract risk type may include at least one of the following categories: contract violation words, risk clause identification, asymmetry category of obligations and rights, and normal category. Each type of annotation is defined by compliance experts based on current laws, regulations and industry standards, and a multi-level review mechanism is used to ensure the accuracy and consistency of the annotation data.

[0027] It can be understood that the classification model in the embodiment of the present invention is a multi-classification model.

[0028] It should be noted that the structure of the dual discriminator generative adversarial network contains the following core components: ① Generator Input: Noise vector.

[0029] Structure: 5-layer fully connected neural network, the activation function is LeakyReLU activation function.

[0030] Function: Map a noise vector to synthetic contract data.

[0031] ②The first discriminator (authenticity discriminator) Input: real contract data or generated data.

[0032] Structure: 4-layer convolutional neural network, the activation function is LeakyReLU activation function, and the output is a scalar probability value.

[0033] Function: Determine whether the input data is real data or generated data.

[0034] ③The second discriminator (category distribution discriminator) Input: real contract data or generated data.

[0035] Structure: The first three convolutional layers are shared with the first discriminator, and the last layer is replaced by a fully connected layer and a Softmax function, and the output is a category probability distribution.

[0036] Function: Predict the risk category distribution of input data and constrain the category rationality of generated data.

[0037] In some embodiments of the present invention, optionally, the training process of the enterprise contract data expansion model includes: 1) Initialize the network parameters of the generator and discriminator of the generative adversarial network to solve the initial weight and bias setting problem of the model and ensure the effectiveness and stability of subsequent training. is the initial parameter of the generator, is the initial parameter of the discriminator, which is initialized randomly and obeys a normal distribution with a mean of 0 and a variance of the unit matrix.

[0038] 2) Adaptively adjust the input noise. The discriminator dynamically corrects the noise vector by monitoring the quality of the enterprise contract data generated in the previous iteration (starting from the second iteration) to solve the problem that the generator is prone to fall into the local optimum and the diversity of enterprise contract data is insufficient. The quality of generated enterprise contract data is continuously improved during the training process, which is expressed as:

[0039] Furthermore, the noise vector is updated according to the gradient, and the calculation method is expressed as:

[0040] In the formula, is the gradient calculation function of the discriminator loss with respect to the noise vector, For the The noise vector at the iteration. The noise vector is a set of random numbers input to the generator in the generative adversarial network, which is used to guide the generator to create diversified synthetic data that is similar to the distribution of real data; are the parameters of the discriminator, For the The noise vector at the iteration, is the learning rate for the noise vector update, is the discriminator function, is the generator function, The loss function for the discriminator to distinguish the authenticity of real and generated enterprise contract data, is the symbol of partial derivative, is a binary classification label (the real enterprise contract data label is 1, and the generated enterprise contract data label is 0). Preferably, Set to 0.3.

[0041] 3) Mix the generated enterprise contract data with the real enterprise contract data. By setting the mixing ratio, we can solve the problem of category imbalance during model training and the problem that a single enterprise contract data source easily leads to model overfitting, achieve more comprehensive and balanced training of the discriminator, and improve the accuracy of the discriminator. The calculation method is expressed as:

[0042] In the formula, For mixed enterprise contract data, Enterprise contract data generated by the generator, For real enterprise contract data, is a mixing ratio parameter. In some embodiments, preferably, Set to 0.3.

[0043] 4) Analyze and verify the topological features of the generated enterprise contract data. By adopting a topological enterprise contract data analysis method, the consistency problem of high-dimensional enterprise contract data in the real topological structure is solved, so that the internal structure of the enterprise contract data is more similar to the real enterprise contract data. The topological enterprise contract data analysis method calculates the topological loss function based on the Wasserstein distance, and calculates the minimum cost required to convert the probability distribution of the enterprise contract data generated by the generator into the probability distribution of the real enterprise contract data, that is, to find the optimal mapping between the enterprise contract data generated by the generator and the distribution of the real enterprise contract data. The calculation method is expressed as:

[0044] Furthermore, the generator is updated in combination with the topological loss, and the update method is expressed as:

[0045] In the formula, is the function for calculating the Wasserstein distance, It is the probability density function, which is used to characterize the distribution of generated data and real data. Specifically, it estimates the probability density of the distribution by performing kernel density estimation (KDE) on the data samples. is the topological loss, To balance the hyperparameters of the general quality and topological quality of the generated enterprise contract data, For the The parameters of the generator at the iteration, For the The parameters of the generator at the iteration, represents the gradient with respect to the generator parameters, is the learning rate updated by the generator, is the authenticity loss function of the generator. Preferably, Set to 0.3, Set to 0.01.

[0046] 5) Two discriminators are used to implement a double discrimination process. The first discriminator is used to judge the authenticity of the enterprise contract data, and the second discriminator is used to judge the category distribution of the enterprise contract data. The KL divergence is used to measure the difference between the true category distribution and the generated enterprise contract data category distribution, so as to solve the problem that the model is difficult to balance between the authenticity and category accuracy of the generated enterprise contract data. The calculation method is expressed as:

[0047] Furthermore, the KL divergence of the category distribution is used to improve the discrimination effect. The calculation method is expressed as:

[0048] In the formula, is the loss function of the discriminator regarding the authenticity of enterprise contract data, is the discriminator balance coefficient, is the loss function of the discriminator for category discrimination, is the total number of categories, is a positive integer, is the true category distribution in The probability of the categories, Predict the category distribution for the discriminator The probability of each category. Preferably, Set to 0.3.

[0049] 6) A feature matching constraint is used in the loss function of the generator to make the generated enterprise contract data consistent with the real enterprise contract data in the learnable feature space, so as to solve the problems such as mode collapse that may be caused by relying solely on the authenticity judgment of the discriminator, and to improve the authenticity and diversity of enterprise contract data at the feature level. The calculation method is expressed as:

[0050] In the formula, is the authenticity loss function of the generator, The middle layer of the generator is used to extract the intermediate features of the enterprise contract data; is the L2 norm, which is used to measure the degree of feature difference.

[0051] 7) Based on the feedback of the dual discriminator, the training strategies of the generator and the discriminator are dynamically optimized to solve the problem that the learning rate and the weight of the loss function are difficult to select during the training process of the model. The weight parameters are dynamically adjusted for different training stages to improve the overall convergence speed and accuracy. The discriminator parameters are updated based on the monitoring of the authenticity and category loss of the discriminator. The calculation method is expressed as:

[0052] In the formula, are the discriminator parameters, is the number of iterations, is the learning rate updated by the discriminator, is the gradient of the loss function with respect to the discriminator parameters, For the The parameters of the discriminator at the iteration, For the The parameters of the discriminator at the iteration. Preferably, Set to 0.3.

[0053] 8) Repeat the above steps until the preset stop iteration condition is met, indicating that the model training is completed. In one embodiment, the preset stop iteration condition is reaching a preset maximum number of iterations. Preferably, the preset maximum number of iterations is set to 1000 times.

[0054] After the enterprise contract data expansion model training is completed, the trained enterprise contract data expansion model is used to increase the number of samples. In some optional embodiments, assuming that the original collected samples are 800, and the enterprise contract data expansion model expands and generates 200 samples, the expanded enterprise contract data set contains 1000 samples.

[0055] The deep learning model unit automatically identifies potential violations in contracts through deep learning models.

[0056] The present invention adopts a 5-layer fully connected neural network (i.e., the neural network model in the embodiment of the present invention) as a deep learning model to perform feature extraction of enterprise contract data (i.e., feature extraction of the first contract sample set in the embodiment of the present invention). Compared with the traditional few-layer (less than 3 layers) neural network design, the multi-layer (more than 3 layers) neural network structure is more conducive to capturing the complex characteristics of enterprise contract data, such as the hierarchy of contract terms and the implicit association between clauses. However, deep networks may encounter problems such as gradient vanishing, gradient exploding, or falling into local optimal solutions, affecting the stability of training and the performance of the model.

[0057] In the prior art, the back propagation neural network is often optimized based on the bat algorithm, and the bat optimization algorithm is used to optimize the neural network parameters, thereby realizing the neural network model to process the enterprise contract data; However, the traditional bat algorithm searches with a single sound wave attribute (such as frequency) as the core, which is difficult to adapt to the optimization requirements of high-dimensional and complex parameter spaces. The present invention proposes a neural network algorithm based on multi-dimensional sound wave bat optimization (i.e., a feature extraction model, which is obtained by training the neural network model of multi-dimensional sound wave bat optimization) to extract features from enterprise contract data. In the training process of the neural network, the multi-dimensional sound wave dynamic modulation method is used to expand the sound wave attributes into four dimensions of frequency, amplitude, wavelength and phase, and these attributes are dynamically adjusted to achieve more refined searches, thereby enhancing the search flexibility of the algorithm in high-dimensional parameter spaces, avoiding the problem that the traditional algorithm may fall into local optimality due to the single search strategy, and effectively improving the accuracy and efficiency of feature extraction when extracting features from high-dimensional and highly heterogeneous samples such as enterprise contract data. In addition, the present invention also uses a global resonance detection mechanism to make up for the lack of perception of the global coupling effect of the traditional bat algorithm, and enhances the optimization ability of the global target by analyzing the relationship between the phase and wavelength between the parameters, further improving the stability and efficiency of the search.

[0058] In some embodiments of the present invention, optionally, the training process of the feature extraction model is as follows: A1. Randomly initialize each parameter and bias in the neural network, and equip each parameter with a multi-dimensional acoustic wave model to support the subsequent dynamic modulation of multi-dimensional attributes, solve the problem of effective coverage of the initial search space, and solve the problem of local optimal traps that may occur in the feature extraction of enterprise contract data. The initialization method is expressed as:

[0059]

[0060] In the formula, is the value of the neural network weight parameter at the initial moment; is the value of the neural network bias parameter at the initial moment; is the variance hyperparameter of the neural network parameters; Indicates compliance with a specific distribution; It means the mean is 0 and the variance is Normal distribution of represents a normal distribution. Preferably, Set to 0.1.

[0061] Furthermore, in order to track and update the multidimensional acoustic wave features of each parameter during the neural network training process, the acoustic wave feature vector is initialized and expressed as:

[0062] In the formula, is the acoustic wave feature vector of the neural network parameters; is the initial frequency; is the initial amplitude, is the initial wavelength, is the initial phase. Preferably, the initial values ​​of the acoustic wave feature vectors of the neural network parameters are all from a mean of 0 and a variance of is obtained from the normal distribution of .

[0063] A2. In the multi-dimensional sound wave modulation link, the sound wave properties of each parameter are dynamically adjusted according to the current performance feedback of the neural network to solve the problem of fine control of the search direction and range by the bat algorithm, which is expressed as:

[0064] In the formula, is the neural network parameter in The frequency at the iteration; is the neural network parameter in The frequency at the iteration; is the partial derivative of the loss function of the neural network with respect to the weight parameter; is the loss function of the neural network; is the frequency regulation rate; is the weight parameter of the neural network; is the neural network parameter in The frequency update increment at the iteration. Preferably, the loss function of the neural network adopts the cross entropy loss, and the class probability is calculated for the enterprise contract data feature vector extracted by the neural network through the preset Softmax function, and then the cross entropy loss function is used to calculate the loss function. The loss function measures the ability of the neural network model to extract the contract data features. For example, when the enterprise contract data feature vector is extracted by the preset Softmax function neural network, the loss function calculates the degree of inconsistency between the extracted features and the actual category labels. During the training process, the optimization algorithm adjusts the model parameters according to the gradient of the loss function, gradually reducing the gap between the prediction and the actual, thereby extracting contract data features that are more in line with semantic requirements.

[0065] Furthermore, the calculation of the frequency adjustment rate depends on the gradient of the current loss function, and the calculation method is expressed as:

[0066] In the formula, is a hyperparameter that controls the slope of the curve. Preferably, Set to 0.1.

[0067] Furthermore, in order to link the amplitude, wavelength and phase of the modulated sound wave so that the global and local exploration balance can be taken into account during the search process, the update method is expressed as:

[0068]

[0069]

[0070] In the formula, is the neural network parameter in The amplitude at the iteration; is the neural network parameter in The amplitude at the iteration; is the neural network parameter in The amplitude update increment at the iteration; is the neural network parameter in The wavelength at the iteration; is the neural network parameter in The wavelength at the iteration; is the neural network parameter in The phase at iteration ; is the neural network parameter in The phase at iteration ; is a symbolic function; is the modulation rate of the amplitude; is the modulation rate of wavelength; is the modulation rate of the phase. Preferably, Set to 0.1, Set to 0.2, Set to 0.1.

[0071] Furthermore, in the multi-dimensional sound wave modulation process, an adaptive spectrum decomposition method is used to dynamically decompose and reconstruct the spectrum characteristics of the sound wave signal to avoid the neural network parameters falling into the local optimum. Based on the decomposition and reconstruction of the sound wave spectrum, a refined search is achieved. The calculation method is expressed as:

[0072] In the formula, is the spectrum decomposition function of the current frequency and amplitude of the neural network parameters; is the number of decomposed sub-bands; For the The frequency of the sub-band, For the The amplitude of each sub-band, is a positive integer; is the frequency parameter of the bat algorithm; is the index of the sub-band; is the amplitude parameter of the bat algorithm.

[0073] Furthermore, the decomposition results are adaptively weighted and integrated to calculate the update increments of frequency and amplitude. The calculation method is expressed as:

[0074]

[0075] In the formula, The frequency parameter is The weight of each sub-band, is a positive integer; The amplitude parameter is The weight of each sub-band.

[0076] Furthermore, the frequency and amplitude parameters have an important influence on the The weights of the sub-bands are used to adaptively weigh The effectiveness of the sub-band in the loss function constraint is calculated as:

[0077]

[0078] In the formula, The frequency parameter is The weight of each sub-band affects the hyperparameter, is a positive integer; The amplitude parameter is The weight of each sub-band affects the hyperparameter; is a natural exponential function. Preferably, Set to 0.1, Set to 0.5.

[0079] A3. In the global resonance detection link, the interaction strength between parameters is evaluated and adjusted so that each parameter can make full use of the global coupling effect brought by phase interaction during the search process, perceive the overall distribution of the parameter set and improve the search stability. The global resonance detection mechanism can make up for the shortcomings of traditional neural network algorithms in the perception of global coupling effects and enhance the collaborative optimization capabilities of neural network parameters. For example, by analyzing the interactions between various features in enterprise contract data, neural networks can more accurately extract the implicit semantics and logical structures in enterprise contract data, ensure search flexibility in high-dimensional complex spaces, and make the model's representation of contract data more refined. The calculation method is expressed as:

[0080] In the formula, is the phase parameter of the bat algorithm; is the resonance adjustment term of the neural network parameters; is the phase parameter of the bat algorithm of the previous iteration; A factor to weight the relative distance of parameter adjustment; is the wavelength parameter of the bat algorithm; is the wavelength parameter of the bat algorithm of the previous iteration. Preferably, the wavelength parameter of the bat algorithm is the same as the index of the iteration number.

[0081] Furthermore, the factor weighting the relative distance of parameter adjustment is adjusted based on the relative position and performance impact of the parameter during the iterative training process, and the calculation method is expressed as:

[0082] In the formula, is the L2 norm; is the weight parameter of the neural network; is the weight parameter of the previous iteration of the neural network; is the attenuation scale parameter. Preferably, Set to 0.95.

[0083] A4. In the echo intensity evaluation phase, the quality of parameter configuration is evaluated by mapping the relationship between the sound wave amplitude and the loss function value, solving the problem of objectively measuring the current search status, and directly confirming the quality of parameters based on the sound wave reflection intensity. The calculation method is expressed as:

[0084] In the formula, is the echo loss; is the neural network parameter in The amplitude at the iteration; Indicates the maximum value among all current amplitudes.

[0085] A5. In the parameter resonance tuning link, the coordination of related parameters is enhanced according to the resonance detection results to make full use of the synergy between parameters and improve the search efficiency, so as to converge quickly in the local area. The calculation method is expressed as:

[0086]

[0087] In the formula, For the neural network The weight parameter at the iteration, For the neural network Bias parameter at iteration; is the first hyperparameter of the tuning step size; is the second hyperparameter for the tuning step size; Update the influence term for the bias parameters of the neural network; is the resonance adjustment term of the neural network parameters; is the global and local exploration function. Preferably, Set to 0.05, Set to 0.3.

[0088] A6. In the dynamic position update and multi-channel exploration phase, the position of the parameters is updated simultaneously in multiple directions, and additional displacement is provided in combination with the characteristics of the sound wave. By calculating the update influence terms of the neural network weights and biases, a full search for potential optimal solutions is achieved in the high-dimensional space, and global and local exploration is taken into account in the complex parameter space. The calculation method is expressed as:

[0089]

[0090] In the formula, is the gradient of the echo loss with respect to the neural network weights; is the gradient of the echo loss to the bias of the neural network; is the learning rate updated based on the echo loss, The step size for updating the acoustic wave characteristics; is the additional update amount for the neural network weights; is the additional update amount of the neural network bias. Preferably, Set to 0.001, Set to 0.01.

[0091] Furthermore, the calculation method of the additional update amount of the neural network weight and the additional update amount of the neural network bias is expressed as:

[0092]

[0093] In the formula, is the step size for this additional displacement, preferably, Set to 0.01. is the amplitude parameter of the bat algorithm; is the wavelength parameter of the bat algorithm; is the phase parameter of the bat algorithm.

[0094] A7. Determine whether the search has reached convergence by monitoring the parameter update amplitude. If the neural network training reaches convergence, stop the iteration. Determine the convergence state of the training by monitoring the parameter update amplitude and set the convergence threshold hyperparameter. Ensure that the model can reach convergence stably and efficiently in the scenario of heterogeneous data and high noise of enterprise contract data. The convergence condition is expressed as:

[0095] In the formula, is the maximum value of all parameter and bias updates; is the convergence threshold hyperparameter, preferably, Set to .

[0096] After the neural network training is completed, the trained neural network is used to extract features of the enterprise contract data. Furthermore, the feature-extracted data is input into the preset Softmax function to calculate the class probability, and the class with the largest class probability is taken as the final prediction category, such as: contract violation words, risk clause identification, asymmetry categories of obligations and rights, and normal categories.

[0097] In an embodiment of the present invention, the enterprise contract data is obtained and input into a pre-trained enterprise contract data expansion model to obtain the expanded contract data; the enterprise contract data and the expanded contract data are combined to obtain a first contract sample set; the pre-trained feature extraction model is used to extract features from the first contract sample set to obtain the contract features of each contract sample in the first contract sample set; the contract features are labeled according to the risk type of the contract sample, and all the labeled contract features are combined to obtain a training feature set; the classification model is trained using the training feature set to obtain an enterprise violation risk prediction model; wherein: the enterprise contract data expansion model is trained based on an adversarial network algorithm model generated based on a dual discriminator; the feature extraction model is trained based on a neural network model optimized by a multi-dimensional sonic bat It is obtained that the enterprise contract data expansion model obtained by training according to the adversarial network algorithm model generated based on the dual discriminator realizes the expansion of the enterprise contract data and increases the scale of training data; and, the feature extraction model obtained by training according to the neural network model optimized by multi-dimensional sonic bats is used to extract features of the first contract sample set, and the training feature set is obtained by annotating based on the extracted contract features, thereby improving the quality of the training data (i.e., the training feature set) used to train the classification model, and then improving the classification model's cognitive accuracy of the sample features. The trained enterprise violation risk prediction model can fully explore the complex semantics and logical associations in the contract terms, and then obtain accurate violation risk prediction results. The present invention can train an enterprise violation risk prediction model that can make high-accuracy judgments on enterprise contract risks.

[0098] In some embodiments of the present invention, optionally, the training process of the enterprise contract data expansion model includes: After the first round of training iterations for the adversarial network algorithm model is completed, in each round of training iterations, the noise vector of the current round of training iterations is corrected according to the contract data generated by the adversarial network algorithm model that completed the previous round of training iterations.

[0099] It should be noted that the noise vector is a set of random numbers input to the generator in the generative adversarial network, which is used to guide the generator to create diverse synthetic data that is similar to the distribution of real data.

[0100] After the first round of training iterations is completed for the adversarial network algorithm model, the present invention performs each round of training iterations, that is, each round of training iterations starting from the second round of training iterations. It can be understood that for the first round of training iterations, there is no previous round of training iterations.

[0101] To illustrate with a specific example, as mentioned above, the training process of the enterprise contract data expansion model includes: 2) Adaptively adjust the input noise and dynamically correct the noise vector by monitoring the quality of the enterprise contract data generated in the previous iteration (starting from the second iteration) to solve the problem that the generator is prone to fall into local optimality and the lack of diversity of enterprise contract data. During the training process, the quality of generated enterprise contract data is continuously improved, which can be expressed as:

[0102] Furthermore, the noise vector is updated according to the gradient, and the calculation method is expressed as:

[0103] In the formula, is the gradient calculation function of the discriminator loss with respect to the noise vector, For the The noise vector at the iteration, are the parameters of the discriminator, For the The noise vector at the iteration, is the learning rate for the noise vector update, is the discriminator function, is the generator function, The loss function for the discriminator to distinguish the authenticity of real and generated enterprise contract data, is a binary classification label (the real enterprise contract data label is 1, and the generated enterprise contract data label is 0). Preferably, Set to 0.3.

[0104] In an embodiment of the present invention, after completing the first round of training iterations for the adversarial network algorithm model, in each round of training iterations, the noise vector of the current round of training iterations is corrected according to the contract data generated by the adversarial network algorithm model that completed the previous round of training iterations, so as to guide the training process, effectively avoid the problem that the generator is prone to falling into local optimality during the training process, and ensure that the enterprise contract data expansion model obtained through training can be expanded to obtain diverse expanded contract data.

[0105] In some embodiments of the present invention, optionally, in the adversarial network algorithm model, one discriminator is used to judge the authenticity of the contract data generated by the adversarial network algorithm model, and another discriminator is used to judge the category distribution of the contract data generated by the adversarial network algorithm model.

[0106] In the embodiment of the present invention, through the above-mentioned adversarial network algorithm model, one discriminator is used to judge the authenticity of the contract data generated by the adversarial network algorithm model, and the other discriminator is used to judge the category distribution of the contract data generated by the adversarial network algorithm model. The embodiment of the present invention can further use KL divergence to measure the difference between the true category distribution and the category distribution of the generated enterprise contract data, so as to solve the problem that the model is difficult to balance the authenticity and category accuracy of the generated enterprise contract data. The calculation method is expressed as:

[0107] Furthermore, the KL divergence of the category distribution is used to improve the discrimination effect. The calculation method is expressed as:

[0108] In the formula, is the loss function of the discriminator regarding the authenticity of enterprise contract data, is the discriminator balance coefficient, is the loss function of the discriminator for category discrimination, is the total number of categories, is a positive integer, is the true category distribution in The probability of the categories, Predict the category distribution for the discriminator The probability of each category. Preferably, Set to 0.3.

[0109] In some embodiments of the present invention, optionally, each pair of adversarial network algorithm models completes one training iteration, verifies whether the total number of current training iterations exceeds a preset iteration number threshold, and obtains a verification result; If the verification result indicates that the total number of current training iterations is less than the iteration number threshold, continue to perform the next round of training iterations on the adversarial network algorithm model; If the verification result indicates that the total number of current training iterations is greater than or equal to the iteration number threshold, the next round of iterative training is stopped, and the adversarial network algorithm model after this training iteration is used as the enterprise contract data expansion model.

[0110] It should be noted that the iteration number threshold is specifically set by the user according to actual needs. In the embodiment of the present invention, by setting the iteration number threshold, and completing one training iteration for each pair of adversarial network algorithm models, verifying whether the total number of current training iterations exceeds the preset iteration number threshold, and obtaining the verification result; if the verification result indicates that the total number of current training iterations is less than the iteration number threshold, continue to perform the next round of training iterations on the adversarial network algorithm model; if the verification result indicates that the total number of current training iterations is greater than or equal to the iteration number threshold, stop the next round of iterative training, and use the adversarial network algorithm model after this training iteration as the enterprise contract data expansion model. The embodiment of the present invention realizes the control of model training iteration, can avoid repeated excessive training, and is conducive to improving training efficiency.

[0111] In some embodiments of the present invention, optionally, the training method of the feature extraction model includes: Each pair of neural network models completes one training iteration, and determines whether the neural network model after this training iteration meets the convergence condition, and obtains the judgment result; If the judgment result indicates that the convergence condition is not met, continue to perform the next round of training iterations on the neural network model; If the judgment result indicates that the convergence condition is met, the next round of iterative training is stopped, and the neural network model after this training iteration is used as the feature extraction model; The expression of the convergence condition is as follows:

[0112] In the formula, It is the maximum value of all parameter and bias updates in the neural network model; is the convergence threshold hyperparameter.

[0113] In the embodiment of the present invention, a training iteration is completed for each pair of neural network models, and it is determined whether the neural network model after this training iteration meets the convergence condition to obtain a judgment result; if the judgment result indicates that the convergence condition is not met, the neural network model continues to be trained for the next round of iterations; if the judgment result indicates that the convergence condition is met, the next round of iterative training is stopped, and the neural network model after this training iteration is used as the feature extraction model; the embodiment of the present invention realizes the control of model training iteration, can avoid excessive repetition of training, and is conducive to improving training efficiency.

[0114] In some embodiments of the present invention, optionally, the training method of the feature extraction model includes: In the multi-dimensional sound wave modulation link, the sound wave properties of each parameter in the neural network are corrected according to the performance of the neural network after the most recent training iteration at the current moment; The expression for the correction processing is as follows:

[0115] In the formula, is the parameter in the neural network The frequency at the iteration; is the parameter in the neural network The frequency at the iteration; is the partial derivative of the loss function of the neural network with respect to the weight parameter; is the loss function of the neural network; is the frequency regulation rate; is the parameter in the neural network The frequency update increment for each iteration.

[0116] Specifically in the embodiment of the present invention, the sound wave attribute includes frequency.

[0117] In some embodiments of the present invention, optionally, the loss function of the neural network adopts cross entropy loss, and the class probability calculation is performed on the enterprise contract data feature vector extracted by the neural network through a preset Softmax function, and then the cross entropy loss function is used to calculate. The loss function measures the ability of the neural network model to extract contract data features. For example, when the enterprise contract data feature vector (i.e., the contract feature in the embodiment of the present invention) is extracted by the preset Softmax function neural network, the loss function calculates the degree of inconsistency between the extracted features and the actual category labels. During the training process, the optimization algorithm adjusts the model parameters according to the gradient of the loss function, gradually reducing the gap between the prediction and the actual, thereby extracting contract data features that are more in line with semantic requirements (i.e., the contract features in the embodiment of the present invention).

[0118] The following is described in conjunction with specific embodiments: See also Figure 2 As shown, Figure 2 The following is a schematic diagram of the principle of a method for predicting and warning corporate compliance violations based on deep learning, where: B1. User interface unit The user interface unit provides an interactive interface between users and the system, which is used for corporate contract upload, violation detection initiation and result viewing. It accepts contract files in multiple formats (such as PDF, Word, TXT). Users can trigger the detection process by clicking a button and call the back-end API to achieve corporate compliance and violation prediction and early warning.

[0119] B2. Data acquisition and preprocessing unit The data collection and preprocessing unit parses, cleans and formats the contract uploaded by the user for subsequent processing and analysis. In one embodiment, the contract content is extracted into text using open source tools (such as PyPDF2 for parsing PDF and python-docx for parsing Word documents).

[0120] In one embodiment, data collection originates from the enterprise's contract document library, which mainly includes various contract texts, attachments, and related documents generated during the contract execution process.

[0121] Furthermore, irrelevant information such as blank lines and special characters are removed from the extracted text, which is specifically achieved using regular expressions and text processing libraries (such as Python's re and nltk).

[0122] Furthermore, the contract content is processed by chapters and sections to provide standardized input for the detection module, segmented according to the contractual format (such as "Article 3" or "Title of Strategic Cooperation Contract Clauses"), and stored as structured data.

[0123] B3. Illegal word library and rule management unit The violation vocabulary and rule management unit maintains the system's violation vocabulary and rule library, and provides basic reference information for model optimization. This unit stores common risk vocabulary (such as "the responsibility for force majeure shall be borne entirely by Party A") and supports hierarchical labeling.

[0124] In one embodiment, as shown in Table 1, the marked categories include: contract violation words, risk clause identification, asymmetry categories of obligations and rights, and normal categories. Each type of marking is defined by compliance experts based on current laws, regulations and industry standards, and a multi-level review mechanism is used to ensure the accuracy and consistency of the marked data.

[0125] Table 1 Comparison table of annotation categories Example of a contract paragraph Label category "Party A shall pay Party B 30% of the total contract amount within three days after the project starts." Identification of risk clauses "Party B has the right to adjust the contract price according to market conditions." Contract Violation Words "Party A and Party B shall jointly bear the losses caused by force majeure." Asymmetric categories of obligations and rights "The contract shall come into effect on the date when both parties sign and seal it." Normal category "If Party B fails to deliver the results as scheduled, it shall pay liquidated damages." Identification of risk clauses Furthermore, the collected data is vectorized. The present invention adopts the Word2Vec algorithm to vectorize the text. The Word2Vec algorithm is a commonly used vectorization algorithm in the field. It scans the texts to be vectorized based on a preset large-scale corpus, and represents each word as a one-hot encoded vector. The dimension of the one-hot encoded vector is equal to the size of the vocabulary in the corpus.

[0126] Furthermore, a database (such as MySQL) is used to store the collected vectorized data and annotation content, and the illegal vocabulary is regularly expanded based on the detection results and user feedback.

[0127] In the task of predicting and warning corporate compliance violations, the rule matching method has insufficient coverage when detecting violations in complex contract texts, and the deep learning model has limited ability to identify fine-grained violation clauses. The present invention combines rule matching based on a violation vocabulary with contract analysis by a deep learning model, giving priority to matching clear rules, and calling deep learning models to predict complex violation clauses that cannot be matched.

[0128] B4, training sample expansion unit It is understandable that in the task of the present invention, the collection, acquisition, labeling and preprocessing of enterprise contract training data are time-consuming and labor-intensive, and insufficient training samples can easily lead to poor generalization ability of the model and affect the accuracy of the model.

[0129] The present invention adopts a generative adversarial network based on a dual discriminator to generate samples, thereby realizing the expansion of enterprise contract data, and improving the diversity, quality and topological consistency of generated enterprise contract data in high-dimensional space.

[0130] Specifically, for the training process of the generative adversarial network algorithm based on the dual discriminator, please refer to the training process of the enterprise contract data expansion model above (steps 1) to 8)), which will not be repeated here.

[0131] Enterprise contract data usually have problems such as insufficient sample size and imbalanced categories, which affect the generalization ability of deep learning models. The present invention adopts a generative adversarial network based on a dual discriminator for sample expansion, and uses a generative adversarial network with a dual discriminator to generate high-quality contract samples. By adjusting the noise vector, the proportion of mixed data, the topological feature consistency optimization and the KL divergence constraint, the authenticity of the generated samples and the rationality of the category distribution are ensured.

[0132] B5. Deep Learning Model Unit The deep learning model unit automatically identifies potential violations in contracts through deep learning models.

[0133] The samples of the training sample expansion unit are input into the deep learning model unit. The present invention adopts a 5-layer fully connected neural network as a deep learning model to extract the features of enterprise contract data.

[0134] Compared with the traditional few-layer (less than 3 layers) neural network design, the multi-layer (more than 3 layers) neural network structure is more conducive to capturing the complex characteristics of corporate contract data, such as the hierarchy of contract terms and the implicit associations between clauses. However, deep networks may encounter problems such as gradient vanishing, gradient exploding, or falling into local optimal solutions, affecting the stability of training and the performance of the model.

[0135] The existing technology optimizes the back-propagation neural network binocular vision calibration based on the bat algorithm, uses the bat optimization algorithm to optimize the neural network parameters, and then realizes the neural network model to process the enterprise contract data; However, the traditional bat algorithm searches with a single sound wave attribute (such as frequency) as the core, which is difficult to adapt to the optimization requirements of high-dimensional and complex parameter spaces. The present invention proposes a neural network algorithm based on multi-dimensional sound wave bat optimization to extract features from enterprise contract data. In the training process of the neural network, the multi-dimensional sound wave dynamic modulation method is used to expand the sound wave attributes into four dimensions: frequency, amplitude, wavelength and phase. These attributes are dynamically adjusted to achieve more refined searches, enhance the search flexibility of the algorithm in high-dimensional parameter spaces, and avoid the problem of falling into local optimality due to the single search strategy in traditional algorithms. When extracting features from high-dimensional and highly heterogeneous samples such as enterprise contract data, the accuracy and efficiency of feature extraction can be effectively improved. In addition, the present invention also uses a global resonance detection mechanism to make up for the lack of perception of the global coupling effect of the traditional bat algorithm, and enhances the optimization ability of the global target by analyzing the relationship between the phase and wavelength between the parameters, further improving the stability and efficiency of the search.

[0136] Specifically, the training process of the neural network algorithm based on multi-dimensional sonic bat optimization refers to the training process of the feature extraction model above (steps A1 to A7), which will not be repeated here.

[0137] The hierarchy of contract terms and the implicit associations between clauses are relatively complex. Traditional optimization algorithms cannot efficiently extract the complex features of contract data. The present invention adopts a multi-dimensional sonic bat optimization algorithm to dynamically adjust the frequency, amplitude, wavelength and phase of the deep neural network parameters used for feature extraction of enterprise contract data, enhance the flexibility of parameter search, and combine global resonance detection and adaptive spectrum decomposition to improve the feature extraction capability of high-dimensional contract data.

[0138] B6. Violation Detection Unit The violation detection unit combines the violation vocabulary and deep learning model to analyze the contract content one by one and mark the violation parts. Specifically, first, the contract text is divided into sentences or short paragraphs as analysis units, and then the sentence segmentation function of nltk or spaCy is used to achieve it.

[0139] During detection, for rule matching, priority is given to matching based on the violation vocabulary and rule library; for content that cannot be matched, the deep learning model is called for prediction.

[0140] The test results are stored in the database to support subsequent query and analysis.

[0141] B7. Result feedback and report generation unit The result feedback and report generation unit feeds back the test results to the user in a clear manner and generates a detailed report.

[0142] In the task of predicting and warning of corporate compliance and violation, contracts are mostly unstructured data. Traditional tools are inconvenient to process and it is difficult to quickly feedback the test results. The present invention proposes a user interface unit that supports uploading contracts in various formats, data structured processing and visual feedback of results. The data preprocessing tool extracts the text in segments and stores it in chapters according to the contract format to generate a clear test report.

[0143] The embodiment of the present invention provides a training device for an enterprise violation risk prediction model, see Figure 3 As shown, Figure 3 The principle block diagram of the training device of the enterprise violation risk prediction model is shown in FIG. The training device 30 of the enterprise violation risk prediction model includes: An expansion module 31 is used to obtain enterprise contract data and input the enterprise contract data into a pre-trained enterprise contract data expansion model to obtain expanded contract data; A combining module 32, configured to combine the enterprise contract data with the extended contract data to obtain a first contract sample set; The feature extraction module 33 is used to extract features from the first contract sample set using a pre-trained feature extraction model to obtain contract features of each contract sample in the first contract sample set; annotate the contract features according to the risk type of the contract sample, and combine all the annotated contract features to obtain a training feature set; A training module 34 is used to train the classification model using a training feature set to obtain an enterprise violation risk prediction model; Among them: the enterprise contract data expansion model is obtained by training based on an adversarial network algorithm model generated based on a dual discriminator; the feature extraction model is obtained by training based on a neural network model optimized by multi-dimensional sonic bats.

[0144] In some embodiments of the present invention, optionally, the expansion module 31 is also used to correct the noise vector of the current round of training iteration in each round of training iteration after completing the first round of training iteration of the adversarial network algorithm model, based on the contract data generated by the adversarial network algorithm model that completed the previous round of training iteration.

[0145] In some embodiments of the present invention, optionally, in the adversarial network algorithm model, one discriminator is used to judge the authenticity of the contract data generated by the adversarial network algorithm model, and another discriminator is used to judge the category distribution of the contract data generated by the adversarial network algorithm model.

[0146] In some embodiments of the present invention, optionally, the expansion module 31 is further used to complete one training iteration for each pair of the adversarial network algorithm model, verify whether the total number of current training iterations exceeds a preset iteration number threshold, and obtain a verification result; The expansion module 31 is also used to continue to perform the next round of training iterations on the adversarial network algorithm model if the verification result indicates that the total number of current training iterations is less than the iteration number threshold; The expansion module 31 is also used to stop the next round of iterative training if the verification result indicates that the total number of current training iterations is greater than or equal to the iteration number threshold, and use the adversarial network algorithm model after this training iteration as the enterprise contract data expansion model.

[0147] In some embodiments of the present invention, optionally, the feature extraction module 33 is further used to complete one training iteration for each pair of the neural network model, determine whether the neural network model after this training iteration meets the convergence condition, and obtain a determination result; The feature extraction module 33 is further configured to continue to perform the next round of training iterations on the neural network model if the judgment result indicates that the convergence condition is not met; The feature extraction module 33 is further configured to stop the next round of iterative training if the judgment result indicates that the convergence condition is satisfied, and use the neural network model after the current training iteration as the feature extraction model; The expression of the convergence condition is as follows:

[0148] In the formula, is the maximum value of all parameter and bias updates in the neural network model; is the convergence threshold hyperparameter.

[0149] In some embodiments of the present invention, optionally, the feature extraction module 33 is further used to correct the acoustic wave attribute of each parameter in the neural network according to the performance of the neural network after the most recent training iteration from the current moment in the multi-dimensional acoustic wave modulation link; The expression of the correction process is as follows:

[0150] In the formula, is the parameter of the neural network in the The frequency at the iteration; is the parameter of the neural network in the The frequency at the iteration; is the partial derivative of the loss function of the neural network with respect to the weight parameter; is the loss function of the neural network; is the frequency regulation rate; is the parameter of the neural network in the The frequency update increment for each iteration.

[0151] The training device of the enterprise violation risk prediction model provided by the embodiment of the present invention can achieve Figure 1 to Figure 2 The various processes implemented by the method embodiment and achieving the same technical effect are not described here to avoid repetition.

[0152] An embodiment of the present invention provides an electronic device 40, see Figure 4 As shown, Figure 4 This is a principle block diagram of an electronic device 40 according to an embodiment of the present invention, including a processor 41, a memory 42, and a program or instruction stored in the memory 42 and executable on the processor 41. When the program or instruction is executed by the processor, the steps in the training method of any enterprise violation risk prediction model according to the present invention are implemented.

[0153] An embodiment of the present invention provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the various processes of the embodiment of the training method of the enterprise violation risk prediction model as described above are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0154] The readable storage medium is, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc. In some examples, the readable storage medium may be a non-transitory readable storage medium.

[0155] An embodiment of the present invention also provides a computer program product, including computer instructions. When the computer instructions are executed by a processor, each process of the training method for the enterprise violation risk prediction model in any of the above-mentioned embodiments can be implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0156] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

Claims

1. A training method for an enterprise violation risk prediction model, characterized in that: include: Acquire enterprise contract data, and input the enterprise contract data into a pre-trained enterprise contract data expansion model to obtain expanded contract data; Combining the enterprise contract data with the extended contract data to obtain a first contract sample set; Using a pre-trained feature extraction model to extract features from the first contract sample set, to obtain contract features of each contract sample in the first contract sample set; According to the risk type of the contract sample, the contract features are marked accordingly, and all the marked contract features are combined to obtain a training feature set; The classification model is trained using the training feature set to obtain an enterprise violation risk prediction model; Wherein: the enterprise contract data expansion model is obtained by training the adversarial network algorithm model generated based on the dual discriminator; The feature extraction model is obtained by training a neural network model optimized by multi-dimensional sonic bats.

2. The training method of the enterprise violation risk prediction model according to claim 1 is characterized in that: The training process of the enterprise contract data expansion model includes: After completing the first round of training iterations for the adversarial network algorithm model, in each round of training iterations, the noise vector of the current round of training iterations is corrected according to the contract data generated by the adversarial network algorithm model that completed the previous round of training iterations.

3. The training method of the enterprise violation risk prediction model according to claim 2 is characterized in that: In the adversarial network algorithm model, one discriminator is used to judge the authenticity of the contract data generated by the adversarial network algorithm model, and the other discriminator is used to judge the category distribution of the contract data generated by the adversarial network algorithm model.

4. The training method of the enterprise violation risk prediction model according to claim 3 is characterized in that: Each time the adversarial network algorithm model completes a training iteration, checking whether the total number of current training iterations exceeds a preset iteration number threshold, and obtaining a verification result; If the verification result indicates that the total number of current training iterations is less than the iteration number threshold, continue to perform the next round of training iterations on the adversarial network algorithm model; If the verification result indicates that the total number of current training iterations is greater than or equal to the iteration number threshold, the next round of iterative training is stopped, and the adversarial network algorithm model after this training iteration is used as the enterprise contract data expansion model.

5. The training method of the enterprise violation risk prediction model according to claim 1 is characterized in that: The training method of the feature extraction model includes: After each training iteration of the neural network model, determine whether the neural network model after the training iteration meets the convergence condition, and obtain a determination result; If the judgment result indicates that the convergence condition is not met, continue to perform the next round of training iteration on the neural network model; If the judgment result indicates that the convergence condition is met, the next round of iterative training is stopped, and the neural network model after the current training iteration is used as the feature extraction model; The expression of the convergence condition is as follows: In the formula, is the maximum value of all parameter and bias updates in the neural network model; is the convergence threshold hyperparameter.

6. The training method of the enterprise violation risk prediction model according to claim 1 is characterized in that: The training method of the feature extraction model includes: In the multi-dimensional sound wave modulation link, the sound wave attribute of each parameter in the neural network is corrected according to the performance of the neural network after the most recent training iteration at the current moment; The expression of the correction process is as follows: In the formula, is the parameter of the neural network in the The frequency at the iteration; is the parameter of the neural network in the The frequency at the iteration; is the partial derivative of the loss function of the neural network with respect to the weight parameter; is the loss function of the neural network; is the frequency regulation rate; is the parameter of the neural network in the The frequency update increment for each iteration.

7. A training device for an enterprise violation risk prediction model, characterized in that: include: An expansion module, used for acquiring enterprise contract data, and inputting the enterprise contract data into a pre-trained enterprise contract data expansion model to obtain expanded contract data; A combination module, used for combining the enterprise contract data and the extended contract data to obtain a first contract sample set; A feature extraction module, configured to extract features from the first contract sample set using a pre-trained feature extraction model to obtain contract features of each contract sample in the first contract sample set; According to the risk type of the contract sample, the contract features are marked accordingly, and all the marked contract features are combined to obtain a training feature set; A training module is used to train the classification model using a training feature set to obtain an enterprise violation risk prediction model; Wherein: the enterprise contract data expansion model is obtained by training the adversarial network algorithm model generated based on the dual discriminator; The feature extraction model is obtained by training a neural network model optimized by multi-dimensional sonic bats.

8. An electronic device, characterized in that: It includes a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction, when executed by the processor, implements the steps in the training method of the enterprise violation risk prediction model as described in any one of claims 1 to 6.

9. A readable storage medium, characterized in that: The readable storage medium stores programs or instructions, which, when executed by a processor, implement the steps in the training method of the enterprise violation risk prediction model as described in any one of claims 1 to 6.

10. A computer program product, characterized in that It comprises computer instructions, which, when executed by a processor, implement the steps of the method for training an enterprise violation risk prediction model as described in any one of claims 1 to 6.

Images