Cipher analysis prediction method and system based on large model

Through the cryptographic analysis and prediction method based on large models, the plain text cryptographic data is preprocessed and feature extraction is used to generate a prediction model, which solves the problem of low efficiency of large-scale data processing and low prediction accuracy in the existing technology. The problem is to achieve efficient and accurate password analysis and prediction.

CN118916889APending Publication Date: 2024-11-08北京中睿天下信息技术有限公司
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410967388.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-18
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

The existing plaintext cryptographic analysis and prediction technology has high computational complexity, low efficiency, and low prediction accuracy when processing large-scale data, making it particularly difficult to process complex and random passwords.

Method used

Using a large model-based cryptographic analysis prediction method, a recurrent neural network, a long and short-term memory network or a transformer model is used to preprocess, train and feature extraction of plaintext password data to generate a prediction model, and use this model to perform new plaintext password data. Forecast and adjustment.

Benefits of technology

Effectively process large-scale data, reduce computing complexity, improve prediction accuracy, and improve analysis and prediction efficiency and reliability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118916889A_ABST
    Figure CN118916889A_ABST
Patent Text Reader

Abstract

The invention discloses a large-model-based cryptographic analysis prediction method and a large-model-based cryptographic analysis prediction system. The method comprises the following steps: firstly, collecting plaintext password data, and preprocessing the plaintext password data; then, carrying out large model training by using the preprocessed plaintext password data; the large model generates a prediction model by learning a plaintext password data mode; and finally, predicting the new plaintext password data according to the generated prediction model. And analyzing the prediction result, and adjusting the prediction model according to the analysis result. The invention provides a plaintext cryptographic analysis prediction technology which can process large-scale data, reduce calculation complexity and improve prediction accuracy.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of machine learning, and in particular to a cryptographic analysis prediction method and system based on a large model. Background Art

[0002] The security of plaintext passwords, that is, passwords that have not been encrypted in any form, mainly depends on the complexity and randomness of the password. However, existing plaintext password analysis and prediction technologies are mainly based on statistics and probability theory. These methods have high computational complexity, low efficiency, and limited prediction accuracy when processing large-scale data.

[0003] First, the processing capacity for large-scale plaintext password data is insufficient. In real life, we often need to process a large amount of password data, which may come from a variety of different sources, such as network services, emails, social media, etc. The scale of these data is usually very large, which may reach hundreds of GB or even TB. However, existing plaintext password analysis and prediction technologies often encounter great difficulties when processing such large-scale data. For example, traditional statistics-based methods, such as frequency analysis, require a lot of calculations for each possible password, which is unrealistic on large-scale data. Similarly, probability-based methods, such as Markov chains, also require a lot of calculations for each possible password, which is also unrealistic on large-scale data.

[0004] Secondly, the existing technology has low prediction accuracy for plaintext passwords. In real life, we often need to predict unknown passwords in order to perform various security-related tasks, such as password cracking, security auditing, etc. However, existing plaintext password analysis and prediction technologies often perform poorly in terms of prediction accuracy. For example, the method based on frequency analysis, although it can reflect the distribution law of characters in the password to a certain extent, its prediction effect is often poor for complex passwords. Similarly, the method based on Markov chain, although it can reflect the transfer law of characters in the password, its prediction effect is also often poor.

[0005] Finally, the existing technology does not adequately handle the complexity and randomness of plaintext passwords. In real life, we often encounter various complex passwords, which may contain a variety of different characters, such as letters, numbers, symbols, etc., and the order of these characters may be highly random. However, the existing plaintext password analysis and prediction technology often encounters great difficulties when dealing with such complex and random passwords. For example, the method based on frequency analysis, although it can reflect the distribution law of characters in the password to a certain extent, its analysis effect is often not good for complex and random passwords. Similarly, the method based on Markov chain, although it can reflect the transfer law of characters in the password, its analysis effect is often not good for complex and random passwords. Summary of the invention

[0006] Based on this, the embodiments of the present application provide a cryptographic analysis prediction method and system based on a large model, which can process large-scale data, reduce computational complexity, and improve prediction accuracy.

[0007] In a first aspect, a cryptographic analysis prediction method based on a large model is provided, the method comprising:

[0008] Collect plaintext password data and pre-process the plaintext password data;

[0009] Using the preprocessed plaintext password data to train a large model; wherein the large model includes a recurrent neural network, a long short-term memory network or a transformer model;

[0010] The large model generates a prediction model by learning the plaintext password data pattern; this includes extracting features from the plaintext password data and then training the model based on the features;

[0011] According to the generated prediction model, new plaintext password data is predicted;

[0012] Analyze the forecast results and adjust the forecast model based on the analysis results.

[0013] Optionally, collecting plaintext password data includes collecting plaintext password data from network services, emails, and social media, specifically:

[0014] Identify and select web services, email service providers, and social media platforms that may collect plaintext password data, and obtain authorization from the data source owner before collecting data;

[0015] Automatically collect data from selected data sources using API interfaces or web crawler technology and store the collected data in a database.

[0016] Optionally, preprocessing the plaintext password data includes performing data cleaning, data conversion, and data labeling operations on the plaintext password data, wherein data cleaning includes removing invalid passwords, data conversion includes converting the passwords into a format that can be processed by the model, and data labeling includes adding labels to the passwords.

[0017] Optionally, use the preprocessed plaintext password data for large model training, including:

[0018] Determine the corresponding deep learning model and prepare the required computing environment, including installing the software framework and hardware resources;

[0019] Load the preprocessed plaintext password data and convert the password sequence into a numerical sequence acceptable to the model according to the input requirements of the model;

[0020] Configure the model's parameters, including the number of layers, number of neurons, learning rate, optimizer, hidden layer dimensions, and attention mechanism.

[0021] Use the training set data to train the model, calculate the output through forward propagation, and adjust the model parameters through the backpropagation algorithm to minimize the loss function. During the training process, adjust the hyperparameters according to the model performance;

[0022] Use the test set data to test the model and evaluate the model's prediction accuracy, precision, and recall indicators to ensure that the model's prediction effect meets expectations.

[0023] Optionally, in extracting features from plain text password data, the features include the length of the password, the types of characters in the password, the frequency of the characters in the password, and the sequence of the characters in the password.

[0024] Optionally, analyzing the prediction results and adjusting the prediction model according to the analysis results include:

[0025] Calculate the prediction accuracy, precision, and recall indicators of the model, and optimize the model based on the calculated indicators.

[0026] In a second aspect, a cryptographic analysis prediction system based on a large model is provided, the system comprising:

[0027] The data collection and processing module is used to collect plain text password data and pre-process the plain text password data;

[0028] A model training module, used to use the pre-processed plaintext password data to train a large model; wherein the large model includes a recurrent neural network, a long short-term memory network or a transformer model;

[0029] The generation module, the large model generates a prediction model by learning the plaintext password data pattern; this includes extracting features from the plaintext password data and then training the model based on the features;

[0030] A prediction module, used to predict new plaintext password data based on the generated prediction model;

[0031] The result analysis feedback module is used to analyze the prediction results and adjust the prediction model according to the analysis results.

[0032] In a third aspect, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the large model-based cryptanalysis prediction method described in any one of the first aspects is implemented.

[0033] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the large model-based cryptanalysis prediction method described in any one of the first aspects is implemented.

[0034] In a fifth aspect, a computer program product is provided, comprising a computer program / instruction, which, when executed by a processor, implements the cryptographic analysis prediction method described in any one of the first aspects.

[0035] The beneficial effects brought by the technical solution provided by the embodiment of the present application include at least:

[0036] (1) Ability to process large-scale data: Due to the use of a large model, the present invention can process large-scale plaintext cryptographic data, thereby improving the efficiency of analysis and prediction. In real life, we often need to process a large amount of cryptographic data, which may come from a variety of different sources, such as network services, emails, social media, etc. The scale of these data is usually very large, which may reach hundreds of GB or even several TB. However, existing plaintext cryptographic analysis and prediction technologies often encounter great difficulties when processing such large-scale data. The present invention can effectively process such large-scale data by using a large model, thereby improving the efficiency of analysis and prediction.

[0037] (2) Reduce computational complexity: Due to the use of a large model, the present invention can effectively reduce computational complexity and improve the speed of analysis and prediction. In real life, we often need to complete password analysis and prediction within a limited time, which requires technology that can quickly process data and output results. However, the existing plaintext password analysis and prediction technology has high computational complexity and low efficiency when processing data. By using a large model, the present invention can effectively reduce computational complexity and improve the speed of analysis and prediction.

[0038] (3) Improve prediction accuracy: Due to the use of a large model, the present invention can effectively improve the prediction accuracy and improve the reliability of analysis and prediction. In real life, we often need to predict unknown passwords in order to perform various security-related tasks, such as password cracking, security auditing, etc. However, existing plaintext password analysis and prediction technologies often perform poorly in terms of prediction accuracy. By using a large model, the present invention can effectively improve the prediction accuracy and improve the reliability of analysis and prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the implementation methods or the description of the prior art. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other implementation drawings can be derived from the provided drawings without creative work.

[0040] Figure 1 A flowchart of the steps of a cryptographic analysis prediction method based on a large model provided in an embodiment of the present application;

[0041] Figure 2 A schematic diagram of the design of a large model provided in an embodiment of the present application;

[0042] Figure 3 A generation diagram of the prediction model provided in the embodiment of the present application;

[0043] Figure 4 A design diagram of a prediction module provided in an embodiment of the present application;

[0044] Figure 5 A flowchart of a cryptographic analysis prediction system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0046] In the description of the present invention, the terms "comprises", "has" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or apparatus comprising a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may also include other steps or units that are not explicitly listed but are inherent to these processes, methods, products or apparatuses, or steps or units added based on further optimization schemes conceived by the present invention.

[0047] In cryptography, a plaintext password is a password that has not been encrypted. Its security mainly depends on the complexity and randomness of the password. Existing plaintext password analysis methods are mainly based on statistics and probability theory, such as methods based on frequency analysis and methods based on Markov chains. The effectiveness of these methods is often limited by the amount of data and processing power.

[0048] Frequency analysis is a statistically based plaintext password analysis method. Its basic idea is that for a given password, the frequency of each character can be counted, and then the possibility of the password can be inferred based on this frequency information. The advantage of this method is that it is simple and easy to understand, and the amount of calculation is relatively small; however, its disadvantage is that for complex passwords, its prediction effect is often poor, and its efficiency is often low when processing large-scale data.

[0049] Markov chain is a plaintext cryptanalysis method based on probability theory. Its basic idea is that for a given password, a Markov chain can be constructed, which reflects the transition probability of characters in the password. Then, the possibility of the password can be inferred through this Markov chain. The advantage of this method is that it can reflect the transition rules of characters in the password, and its prediction effect may be better for some specific passwords; however, its disadvantage is that the computational complexity is high, it is not suitable for processing large-scale data, and its prediction effect is often poor for complex passwords.

[0050] The most similar implementation scheme of the present invention may be a plaintext password analysis prediction technology based on deep learning. Deep learning is a machine learning method based on neural networks, which can automatically learn and extract the features of data, and then make predictions based on these features. The advantage of this method is that it can process large-scale data, and its prediction effect may be better for complex passwords; however, its disadvantage is that it requires a large amount of computing resources and a large amount of training data.

[0051] Existing plaintext password analysis methods have high computational complexity, low efficiency, and limited prediction accuracy when processing large-scale data. For example, although the frequency analysis method can reflect the distribution law of characters in the password to a certain extent, its prediction effect is often poor for complex passwords. Although the Markov chain-based method can reflect the transfer law of characters in the password, its computational complexity is high and it is not suitable for processing large-scale data.

[0052] Although the plaintext cryptanalysis prediction technology based on deep learning has solved the above problems to a certain extent, it also has its own problems. First, deep learning requires a lot of computing resources, which may not be realistic in some resource-limited environments. Second, deep learning requires a lot of training data, which may not be realistic in some data-scarce environments. Finally, deep learning models are often black box models, and their internal working mechanisms are often difficult to understand and explain, which may be a problem in some applications that need to understand and explain the prediction results.

[0053] The purpose of the present invention is to provide a large-model-based plaintext cryptanalysis prediction system and method, which can process large-scale data, reduce computational complexity, and improve prediction accuracy. Compared with the existing technology, the present invention has the following advantages:

[0054] (1) Processing large-scale data: Due to the use of a large model, the present invention can process large-scale plaintext and cryptographic data, thereby improving the efficiency of analysis and prediction.

[0055] (2) Reducing computational complexity: Due to the use of a large model, the present invention can effectively reduce computational complexity and improve the speed of analysis and prediction.

[0056] (3) Improving prediction accuracy: Due to the use of a large model, the present invention can effectively improve the prediction accuracy and enhance the reliability of analysis and prediction.

[0057] The plaintext cryptanalysis prediction method provided by the present invention is implemented in a cryptanalysis prediction system, and includes: a large model for receiving and processing plaintext cryptanalysis data, generating a prediction model by learning the pattern of the plaintext cryptanalysis data; and a prediction module for predicting new plaintext cryptanalysis data based on the generated prediction model.

[0058] Please refer to Figure 1 , which shows a flowchart of a cryptographic analysis prediction method based on a large model provided in an embodiment of the present application, the method may include the following steps:

[0059] S1, collect plaintext password data and preprocess the plaintext password data.

[0060] Specifically, identify and select web services, email service providers, and social media platforms that can collect plaintext password data, and obtain authorization from the data source owner before collecting data;

[0061] Automatically collect data from selected data sources using API interfaces or web crawler technology and store the collected data in a database.

[0062] S2, uses the preprocessed plaintext password data for large model training.

[0063] Among them, large models include recurrent neural networks, long short-term memory networks or transformer models.

[0064] In an embodiment of the present application, this step includes the design and implementation of a large model: the large model is based on a deep learning method and can use models such as a recurrent neural network (RNN), a long short-term memory network (LSTM), or a transformer (Transformer). These models can process sequence data and are suitable for processing cryptographic data. The design of a large model needs to take into account factors such as the complexity of the model, the limitations of computing resources, and the characteristics of the data. For example, for complex passwords, more complex models such as LSTM or Transformer are required; for environments with limited computing resources, simpler models such as RNN are required. The implementation of a large model needs to take into account processes such as model training, verification, and testing. For example, model training requires a lot of computing resources and time, and requires the use of high-performance hardware such as GPU, and efficient software such as TensorFlow or PyTorch; model verification and testing require representative data, and data needs to be collected from a variety of different sources. Figure 2 , a design schematic diagram of the large model is given.

[0065] In an optional embodiment of the present application, the pre-processed plaintext password data is used for large model training, including:

[0066] Determine the corresponding deep learning model and prepare the required computing environment, including installing the software framework and hardware resources;

[0067] Load the preprocessed plaintext password data and convert the password sequence into a numerical sequence acceptable to the model according to the input requirements of the model;

[0068] Configure the model's parameters, including the number of layers, number of neurons, learning rate, optimizer, hidden layer dimensions, and attention mechanism.

[0069] Use the training set data to train the model, calculate the output through forward propagation, and adjust the model parameters through the backpropagation algorithm to minimize the loss function. During the training process, adjust the hyperparameters according to the model performance;

[0070] Use the test set data to test the model and evaluate the model's prediction accuracy, precision, and recall indicators to ensure that the model's prediction effect meets expectations.

[0071] S3, the large model generates a prediction model by learning the plaintext password data pattern.

[0072] This includes extracting features from plaintext password data and then training a model based on the features.

[0073] This step includes the generation of a prediction model: the large model generates a prediction model by learning the pattern of plaintext password data. This includes feature extraction of plaintext password data and then training the model based on these features. Feature extraction is a key step that determines whether the model can effectively learn the pattern of the data. Features include the length of the password, the types of characters in the password, the frequency of characters in the password, the sequence of characters in the password, etc. The selection of features needs to take into account factors such as the characteristics of the data, the capabilities of the model, and the prediction goals. For example, for complex passwords, more features need to be extracted, such as the sequence of characters in the password; for simple passwords, only a small number of features need to be extracted, such as the length of the password. Model training needs to take into account factors such as the complexity of the model, the amount of data, and the training goals. For example, for complex models, more data and longer training time are required; for simple models, only a small amount of data and a shorter training time are required. As Figure 3 , the generated graph of the prediction model is given.

[0074] S4, predicting new plaintext password data based on the generated prediction model.

[0075] In this step, the prediction module is designed and implemented: the prediction module is used to predict new plaintext password data based on the generated prediction model. This includes inputting new plaintext password data and then outputting the prediction results based on the prediction model. The design of the prediction module needs to take into account factors such as the module interface, the module functions, and the module performance. For example, the module interface needs to be simple and easy to use, and can easily receive input data and output prediction results; the module functions need to be complete, able to process a variety of input data, and output a variety of prediction results; the module performance needs to be efficient and able to complete the prediction within a limited time. The implementation of the prediction module needs to take into account factors such as the stability of the module, the compatibility of the module, and the maintainability of the module. For example, the stability of the module needs to ensure that the prediction results can be stably output under a variety of input data; the compatibility of the module needs to ensure that it can operate normally in a variety of environments; the maintainability of the module needs to ensure that when problems occur in the module, it can be easily maintained and repaired. Figure 4 The design diagram of the prediction module is given.

[0076] S5, analyzing the prediction results and adjusting the prediction model according to the analysis results.

[0077] like Figure 5 , a flowchart of the cryptanalysis prediction system is given, and the system flow is as follows:

[0078] (1) Data collection and preprocessing: First, a large amount of plaintext password data needs to be collected. This data can come from a variety of different sources, such as network services, emails, social media, etc. After collecting the data, the data needs to be preprocessed to facilitate model training. Preprocessing includes steps such as data cleaning (for example, removing invalid passwords), data conversion (for example, converting passwords into a format that the model can process), and data annotation (for example, adding labels to passwords for supervised learning).

[0079] (2) Training of the big model: After collecting and preprocessing the data, use the data to train the big model. The big model is based on deep learning methods and can use models such as recurrent neural networks (RNN), long short-term memory networks (LSTM), or transformers. During the training process, the model learns the patterns of the password data and generates a prediction model based on these patterns.

[0080] (3) Generation of prediction model: The big model generates a prediction model by learning the patterns of plaintext password data. This includes extracting features from the plaintext password data and then training the model based on these features. Features include the length of the password, the type of characters in the password, the frequency of characters in the password, the sequence of characters in the password, etc.

[0081] (4) Design and implementation of the prediction module: The prediction module is used to predict new plaintext password data based on the generated prediction model. This includes inputting new plaintext password data and then outputting the prediction result based on the prediction model.

[0082] (5) Result analysis and feedback: Finally, the prediction results need to be analyzed to understand the performance of the model and adjust the model based on the analysis results. The prediction accuracy, precision, recall and other indicators of the model can be calculated, and then the model can be optimized based on these indicators.

[0083] The key points of the present invention include: the design and implementation of the large model, the generation of the prediction model, and the design and implementation of the prediction module. The points to be protected include: the specific implementation method of the large model, the generation algorithm of the prediction model, and the specific implementation method of the prediction module.

[0084] The design and implementation of the large model is a key point of the present invention, which determines whether the system can effectively process large-scale plaintext password data. When designing a large model, factors such as the complexity of the model, the limitation of computing resources, and the characteristics of the data need to be considered. When implementing a large model, processes such as model training, verification, and testing need to be considered. Therefore, the design and implementation method of the large model is a point to be protected in the present invention.

[0085] The generation of the prediction model is a key point of the present invention, which determines whether the system can effectively learn the pattern of the plaintext password data. When generating the prediction model, it may be necessary to extract features from the plaintext password data and then train the model based on these features. Feature extraction is a key step that determines whether the model can effectively learn the pattern of the data. Model training is a key step that determines whether the model can effectively predict new plaintext password data. Therefore, the generation algorithm of the prediction model is a point to be protected in the present invention.

[0086] The design and implementation of the prediction module is a key point of the present invention, which determines whether the system can effectively predict new plaintext password data. When designing the prediction module, factors such as the interface of the module, the function of the module, and the performance of the module need to be considered. When implementing the prediction module, factors such as the stability of the module, the compatibility of the module, and the maintainability of the module need to be considered. Therefore, the design and implementation method of the prediction module is a point to be protected in the present invention.

[0087] In the above embodiments of the present application, alternative technical solutions may optionally include using other types of machine learning models, such as neural networks, support vector machines, etc., or using more complex statistical methods for plaintext cryptanalysis and prediction.

[0088] A neural network is a machine learning model based on the working principle of human brain neurons. It can automatically learn and extract the features of data, and then make predictions based on these features. The advantage of a neural network is that it can process large amounts of data, and its prediction effect may be better for complex passwords; however, its disadvantage is that it requires a lot of computing resources and a lot of training data.

[0089] Support vector machine is a machine learning model based on statistics. It divides data into two categories by finding an optimal hyperplane, and then makes predictions based on this hyperplane. The advantage of support vector machine is that the computational complexity is relatively low, and its prediction effect may be good for simple passwords; however, its disadvantage is that its prediction effect may not be good for complex passwords, and its efficiency may be low when processing large-scale data.

[0090] More complex statistical methods may include methods based on Bayesian networks, methods based on decision trees, etc. The advantage of these methods is that they can reflect the distribution pattern of characters in passwords, and their prediction effect may be better for some specific passwords; however, their disadvantage is that they have high computational complexity and are not suitable for processing large-scale data, and their prediction effect may be poor for complex passwords.

[0091] In one embodiment of the present application, a cryptographic analysis prediction system based on a large model is also provided, characterized in that the system includes:

[0092] The data collection and processing module is used to collect plain text password data and pre-process the plain text password data;

[0093] A model training module is used to train a large model using pre-processed plaintext password data; wherein the large model includes a recurrent neural network, a long short-term memory network or a transformer model;

[0094] The generation module, the large model generates a prediction model by learning the plaintext password data pattern; this includes extracting features from the plaintext password data and then training the model based on the features;

[0095] A prediction module, used to predict new plaintext password data based on the generated prediction model;

[0096] The result analysis feedback module is used to analyze the prediction results and adjust the prediction model according to the analysis results.

[0097] For the specific definition of the cryptanalysis prediction system based on the large model, please refer to the definition of the cryptanalysis prediction method based on the large model above, which will not be repeated here. Each module in the above-mentioned cryptanalysis prediction system based on the large model can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0098] In one embodiment, an electronic device is provided, which may be a computer. The electronic device includes a processor, a memory, and a network interface connected via a system bus. The processor of the device is used to provide computing and control capabilities. The memory of the device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used for predicting data based on cryptographic analysis of a large model. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a cryptographic analysis prediction method based on a large model is implemented.

[0099] In one embodiment of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned large-model-based cryptographic analysis prediction method are implemented.

[0100] In one embodiment of the present application, a computer program product is provided, including a computer program / instruction, which implements the steps of the above-mentioned large model-based cryptanalysis prediction method when executed by a processor.

[0101] The computer-readable storage medium and computer program product provided in this embodiment have similar implementation principles and technical effects to those of the above method embodiments, and are not described in detail here.

[0102] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing related hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in M ​​forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (SyMchliMk) DRAM (SLDRAM), memory bus (RaMbus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0103] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0104] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the patent application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent application shall be subject to the attached claims.

Claims

1. A cryptographic analysis prediction method based on a large model, characterized in that: The method comprises: Collect plaintext password data and pre-process the plaintext password data; Using the preprocessed plaintext password data to train a large model; wherein the large model includes a recurrent neural network, a long short-term memory network or a transformer model; The large model generates a prediction model by learning the plaintext password data pattern; this includes extracting features from the plaintext password data and then training the model based on the features; According to the generated prediction model, new plaintext password data is predicted; Analyze the forecast results and adjust the forecast model based on the analysis results.

2. The cryptographic analysis prediction method according to claim 1, characterized in that: Collecting plaintext password data includes collecting plaintext password data from network services, emails, and social media, specifically: Identify and select web services, email service providers, and social media platforms that may collect plaintext password data, and obtain authorization from the data source owner before collecting data; Automatically collect data from selected data sources using API interfaces or web crawler technology and store the collected data in a database.

3. The cryptographic analysis prediction method according to claim 1, characterized in that: Preprocessing the plaintext password data includes data cleaning, data conversion, and data labeling operations on the plaintext password data, wherein data cleaning includes removing invalid passwords, data conversion includes converting passwords into a format that can be processed by the model, and data labeling is adding labels to the passwords.

4. The cryptographic analysis prediction method according to claim 1, characterized in that: Use preprocessed plaintext password data for large model training, including: Determine the corresponding deep learning model and prepare the required computing environment, including installing the software framework and hardware resources; Load the preprocessed plaintext password data and convert the password sequence into a numerical sequence acceptable to the model according to the input requirements of the model; Configure the model's parameters, including the number of layers, number of neurons, learning rate, optimizer, hidden layer dimensions, and attention mechanism. Use the training set data to train the model, calculate the output through forward propagation, and adjust the model parameters through the backpropagation algorithm to minimize the loss function. During the training process, adjust the hyperparameters according to the model performance; Use the test set data to test the model and evaluate the model's prediction accuracy, precision, and recall indicators to ensure that the model's prediction effect meets expectations.

5. The cryptographic analysis prediction method according to claim 1, characterized in that: In the feature extraction of plain text password data, the features include the length of the password, the types of characters in the password, the frequency of the characters in the password, and the sequence of the characters in the password.

6. The cryptographic analysis prediction method according to claim 1, characterized in that: Analyze the forecast results and adjust the forecast model based on the analysis results, including: Calculate the prediction accuracy, precision, and recall indicators of the model, and optimize the model based on the calculated indicators.

7. A cryptographic analysis prediction system based on a large model, characterized in that: The system comprises: The data collection and processing module is used to collect plain text password data and pre-process the plain text password data; A model training module, used to use the pre-processed plaintext password data to train a large model; wherein the large model includes a recurrent neural network, a long short-term memory network or a transformer model; The generation module, the large model generates a prediction model by learning the plaintext password data pattern; this includes extracting features from the plaintext password data and then training the model based on the features; A prediction module, used to predict new plaintext password data based on the generated prediction model; The result analysis feedback module is used to analyze the prediction results and adjust the prediction model according to the analysis results.

8. An electronic device, characterized in that: The invention comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the cryptographic analysis prediction method according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor, the cryptographic analysis prediction method according to any one of claims 1 to 6 is implemented.

10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the cryptographic analysis prediction method as described in any one of claims 1 to 6 is implemented.