Universal machine learning hyper-parameter black box optimization system and method

Abstract

The invention belongs to the technical field of hyper-parameter optimization, and provides a universal machine learning hyper-parameter black box optimization system and method, and the system comprises a hyper-parameter optimization client which is used for submitting an optimization request; the hyper-parameter optimization service system is used for processing a received client request; the optimization algorithm configuration component is used for configuring a hyper-parameter optimization algorithm; the automatic stopping component is used for stopping the meaningless training set; the hyper-parameter optimization control panel is used for management configuration, scheduling monitoring, user management, authority management and system setting; according to the method, the optimal hyper-parameter combination can be quickly found, and the hyper-parameter tuning efficiency and accuracy are remarkably improved; compared with a traditional manual tuning method, the method has the advantages that time and labor cost are greatly saved, interference of human factors is avoided, and the tuning result is more objective and reliable.

Description

Technical Field

[0001] The present invention belongs to the technical field of hyperparameter optimization, and specifically relates to a universal machine learning hyperparameter black box optimization system and method. Background Art

[0002] Machine learning is one of the core technologies of artificial intelligence. It can give artificial intelligence higher intelligence and autonomy, and has been widely used in many fields and industries. The performance and effect of machine learning algorithms are affected by many factors, among which choosing appropriate hyperparameters is one of the key factors.

[0003] In machine learning, the selection of hyperparameters has an important impact on the performance of the model; however, the selection of hyperparameters is often a complex process that requires a lot of experiments and tuning; traditional hyperparameter tuning methods usually require manual intervention, are inefficient, and are easily affected by human factors; in order to solve these problems, many hyperparameter optimization methods have been proposed, such as random search, simulated annealing, particle swarm optimization, genetic algorithms, etc.; although these methods can automatically search and optimize hyperparameters, improving the efficiency and accuracy of hyperparameter tuning; however, these methods still have some problems; for example, the above often requires a lot of computing resources and time, and for some complex models and data sets, it is not easy to find the optimal hyperparameters, which reduces the efficiency and accuracy of hyperparameter tuning.

[0004] Therefore, there is a need for a universal machine learning hyperparameter black-box optimization system and method that can automatically optimize the hyperparameters of various machine learning models and improve the efficiency and accuracy of hyperparameter tuning. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides a universal machine learning hyperparameter black-box optimization system and method to solve the problems in the prior art that the technology often requires a large amount of computing resources and time, and for some complex models and data sets, it is impossible to find the optimal hyperparameters, which reduces the efficiency and accuracy of hyperparameter tuning.

[0006] A general machine learning hyperparameter black box optimization system, including:

[0007] Hyperparameter optimization client, used to submit optimization requests;

[0008] Hyperparameter optimization service system, used to process received client requests;

[0009] Optimization algorithm configuration component, used to configure the hyperparameter optimization algorithm;

[0010] Automatic stopping component, used to stop meaningless training sets;

[0011] Hyperparameter optimization control panel, used to manage configuration, scheduling monitoring, user management, permission management, and system settings.

[0012] Preferably, the hyper-parameter optimization client also includes a function for the user to encapsulate the requests that need to be processed, and the encapsulated information includes machine learning metadata and the resources involved.

[0013] Preferably, the hyper-parameter optimization service system includes:

[0014] Data collection module, used to collect training data and test data for machine learning models;

[0015] Hyperparameter optimization module, used to optimize the hyperparameters of machine learning models;

[0016] Model training module, used to train machine learning models using optimized hyperparameters;

[0017] Model evaluation module, used to evaluate the trained machine learning model;

[0018] The result output module is used to output the optimized hyperparameters and evaluation results.

[0019] Preferably, the hyperparameter optimization module optimizes the hyperparameters of the machine learning model based on random search, simulated annealing, particle swarm optimization, and genetic algorithm.

[0020] Preferably, the model evaluation module uses indicators such as accuracy, recall rate, and F1 value to evaluate the machine learning model.

[0021] Preferably, the result output module outputs the optimized hyperparameters and evaluation results in the form of tables, charts, etc.

[0022] Preferably, the optimization algorithm configuration component includes two types: built-in optimization algorithms and user-defined optimization algorithms, wherein the built-in algorithms include Bayesian optimization, Gaussian process, grid search, random search, etc.

[0023] A general black-box optimization method for hyperparameters in machine learning, comprising the following steps:

[0024] S1. Collect training data and test data for machine learning models;

[0025] S2. Optimize the hyperparameters of the machine learning model;

[0026] S3. Use the optimized hyperparameters to train the machine learning model.

[0027] S4. Evaluate the trained machine learning model;

[0028] S5. Output the optimized hyperparameters and evaluation results.

[0029] A processor is configured to execute the general machine learning hyperparameter black-box optimization system described above.

[0030] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the above-mentioned general machine learning hyperparameter black-box optimization system.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] 1. The present invention adopts automated hyperparameter optimization methods, such as random search, simulated annealing, particle swarm optimization, genetic algorithm, etc., which can quickly find the optimal hyperparameter combination and significantly improve the efficiency and accuracy of hyperparameter tuning. Compared with the traditional manual tuning method, this method greatly saves time and labor costs, while avoiding the interference of human factors, making the tuning results more objective and reliable.

[0033] 2. The system and method of the present invention are applicable to various machine learning models and data sets, and have strong versatility and scalability. Regardless of the model type and size, the system can be used for hyperparameter optimization. At the same time, with the continuous development of machine learning technology, new optimization algorithms and strategies can also be easily integrated into the system to meet the ever-changing needs. The system has built-in state-of-the-art hyperparameter optimization algorithms, such as Bayesian optimization, Gaussian process, grid search, etc. These algorithms can be directly put into production without additional development and debugging.

[0034] 3. The present invention lowers the threshold for machine learning parameter adjustment, so that non-expert users can also perform efficient hyper-parameter optimization; users do not need to pay attention to the specific algorithm implementation and context of the hyper-parameter server, and only need to abide by certain conventions to perform general hyper-parameter optimization work;

[0035] 4. This system supports large-scale applications. It only needs to expand the training nodes on the hyperparameter server side, and no adjustments are required on the user side. This high availability and high scalability enable the system to cope with various complex application scenarios.

[0036] 5. This system has a wider range of applications and is suitable for a variety of machine learning platforms, providing users with more choices and flexibility. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 This is a framework diagram of the machine learning hyperparameter black box optimization system of the present invention;

[0038] Figure 2 This is a framework diagram of the hyper-parameter optimization service system of the present invention;

[0039] Figure 3 This is a flow chart of the general machine learning hyperparameter black box optimization method of the present invention. DETAILED DESCRIPTION

[0040] The following embodiments of the present invention are described in further detail in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0041] Embodiment 1: The present invention provides a general machine learning hyperparameter black box optimization system, such as Figure 1 As shown, including:

[0042] A hyperparameter optimization client, used to submit optimization requests; the hyperparameter optimization client also includes a function for a user to encapsulate a request to be processed, and the encapsulated information includes machine learning meta-information and the resources involved;

[0043] Hyperparameter optimization service system, used to process received client requests;

[0044] The optimization algorithm configuration component is used to configure the hyperparameter optimization algorithm; the optimization algorithm configuration component includes two types: built-in optimization algorithm and user-defined optimization algorithm, where the built-in algorithm includes Bayesian optimization, Gaussian process, grid search, and random search; the optimization algorithm is a conventional technical solution of the prior art and will not be described in detail here;

[0045] Automatic stopping component, used to stop meaningless training sets;

[0046] Hyperparameter optimization control panel, used to manage configuration, scheduling monitoring, user management, permission management, and system settings.

[0047] Specifically, Figure 2 As shown, the hyper-parameter optimization service system includes:

[0048] The data acquisition module is used to collect training data and test data for the machine learning model. The data acquisition module uses an adaptive sampling algorithm, and its algorithm formula is as follows:

[0049]

[0050] where r i is the sampling rate of the ith sample, p i is the density estimate of the area where the sample is located, and α is the adjustment parameter;

[0051] Hyperparameter optimization module, used to optimize the hyperparameters of machine learning models;

[0052] The model training module is used to train the machine learning model using optimized hyperparameters; the specific steps are as follows:

[0053] Step 1: Initialize and distribute the global model: The central server initializes the model and then sets the global model parameters ω g Distribute to each node;

[0054] Step 2: Node local training: The participating nodes receive the global model parameters, use the local data set for local training, and update the parameters of the local model. The calculation method is shown in the following formula; in each training iteration, the stochastic gradient descent method is used as the local optimizer. After the local node model is calculated by the set round of SGD, the new model parameters are uploaded to the central server:

[0055]

[0056] Among them, t represents the current training round, η represents the learning rate, is the gradient calculation symbol, f i (ω k ) is the training instance input parameter ω k losses incurred;

[0057] Step 3: The central server receives the node model parameters for aggregation: The aggregation method is shown in the following formula. When the penalty regularization term u is 1, the corresponding algorithm process is the FedAvg process:

[0058]

[0059] Step 4: Global model distribution: The central server distributes the global model parameters to each node;

[0060] Step 5: Repeat steps 2-4 until convergence or the required number of training rounds is reached;

[0061] The model evaluation module is used to evaluate the trained machine learning model; the model evaluation module uses accuracy, precision, recall, and F1 value indicators to evaluate the machine learning model; the accuracy calculation formula is as follows:

[0062]

[0063] The accuracy calculation formula is as follows:

[0064]

[0065] The recall rate calculation formula is as follows:

[0066]

[0067] The F1 value calculation formula is as follows:

[0068]

[0069] Among them, TP (True Positive) represents the number of samples correctly predicted as positive, TN (True Negative) represents the number of samples correctly predicted as negative, FP (False Positive) represents the number of samples incorrectly predicted as positive, and FN (False Negative) represents the number of samples incorrectly predicted as negative.

[0070] The result output module is used to output the optimized hyperparameters and evaluation results; the result output module outputs the optimized hyperparameters and evaluation results in the form of tables and charts.

[0071] More specifically, the hyperparameter optimization module optimizes the hyperparameters of the machine learning model based on random search, simulated annealing, particle swarm optimization, and genetic algorithm;

[0072] The random search includes:

[0073] Step 1, initialization: set the initial conditions such as the parameter range and search times of random search;

[0074] Step 2: Random sampling: randomly generate a set of hyperparameters within the set parameter range;

[0075] Step 3: Model training and evaluation: Use this set of hyperparameters to train the machine learning model and evaluate its performance (such as accuracy, precision, recall, F1 value, etc.);

[0076] Step 4: Record and compare: Record the current optimal hyperparameter combination and the corresponding model performance, and compare with the previous search results;

[0077] Step 5, Iteration: Repeat steps 2-4 until the set number of searches is reached or a hyperparameter combination that meets the performance requirements is found;

[0078] The search process is as follows: randomly select a set of hyperparameters θ in the hyperparameter space Θ; use this set of hyperparameters to train the machine learning model and evaluate its performance; repeat the above process until the hyperparameters that meet the conditions are found;

[0079] The simulated annealing comprises:

[0080] Step 1, initialization: set the initial temperature, cooling rate, initial hyperparameter combination, etc.;

[0081] Step 2: Current state evaluation: Calculate the performance of the model under the current hyperparameter combination;

[0082] Step 3, neighborhood search: Generate a set of new hyperparameter combinations (i.e., neighborhood solutions) near the current hyperparameter combination;

[0083] Step 4, acceptance probability: According to the probability acceptance criterion of the simulated annealing algorithm, decide whether to accept the new hyperparameter combination; usually, if the performance of the new combination is better, it is accepted; if the performance is worse, it is accepted with a certain probability, and the probability decreases as the temperature decreases;

[0084] Step 5, cooling: lower the temperature according to the set cooling rate;

[0085] Step 6, Iteration: Repeat steps 2-5 until the temperature drops below the set threshold or a hyperparameter combination that meets performance requirements is found;

[0086] Among them, the simulated annealing algorithm includes:

[0087] 1) Assume current temperature T, initial temperature T0, cooling rate α;

[0088] 2) Let the current state be x, and the initial state be x0;

[0089] 3) The new state x′ is generated by random perturbation of the current state x;

[0090] 4) Acceptance probability; where ΔE = E(x′) - E(x) is the energy difference between the new state and the current state;

[0091] The particle swarm optimization comprises:

[0092] Step 1, initialization: set the number of particle swarms, the initial position and velocity of each particle, inertia weight, learning factor, etc.;

[0093] Step 2, fitness evaluation: calculate the performance of the model at the current position of each particle (i.e., hyperparameter combination) as the fitness value;

[0094] Step 3, update individual optimality: compare the current fitness value of each particle with its historical optimal fitness value, and update the individual optimal position;

[0095] Step 4: Update the global optimum: compare the current fitness values ​​of all particles and update the global optimal position;

[0096] Step 5, speed update: update the speed of each particle according to the inertia weight, learning factor, and individual optimal and global optimal positions;

[0097] Step 6, position update: update the position of each particle (i.e., hyperparameter combination) according to the updated velocity;

[0098] Step 7, Iteration: Repeat steps 2-6 until the set number of iterations is reached or a hyperparameter combination that meets the performance requirements is found;

[0099] Wherein, the particle swarm optimization algorithm includes:

[0100] 1) Let the particle position x i and speed v i ;

[0101] 2) Let the personal best position p and the global best position g;

[0102] 3) Its update speed formula:

[0103] 4) Update position formula:

[0104] Among them, w is the inertia weight, c1 and c2 are learning factors, r1 and r2 are random numbers;

[0105] The genetic algorithm comprises:

[0106] Step 1: Initialize the population: randomly generate a set of hyperparameter combinations as the initial population;

[0107] Step 2, fitness evaluation: calculate the fitness value of each individual (i.e., hyperparameter combination);

[0108] Step 3, selection: select a part of individuals as parents according to their fitness values, usually using methods such as roulette selection and tournament selection;

[0109] Step 4, crossover: Perform a crossover operation on the selected parent individuals to generate new offspring individuals (i.e., new hyperparameter combinations); the crossover operation can be performed using single-point crossover, double-point crossover, uniform crossover, etc.

[0110] Step 5, mutation: Perform mutation operations on the newly generated offspring individuals to increase the diversity of the population; the mutation operation can randomly change the values ​​of certain hyperparameters;

[0111] Step 6: Generate a new population: The offspring individuals obtained after crossover and mutation and the unselected individuals in the parent generation form a new population;

[0112] Step 7, Iteration: Repeat steps 2-6 until the set number of iterations is reached or a hyperparameter combination that meets the performance requirements is found;

[0113] Among them, genetic algorithms include:

[0114] 1) Selection operation: select excellent individuals according to the fitness function;

[0115] 2) Crossover operation: exchange some genes of two individuals to generate a new individual;

[0116] 3) Mutation operation: Randomly change the genes of individuals with a certain probability

[0117] Among them, the fitness function can be expressed as f(x), where x It is the genetic code of an individual.

[0118] From the above, we can see that the optimization effects of the hyperparameters of the machine learning model based on random search, simulated annealing, particle swarm optimization, and genetic algorithm are compared as shown in the following table:

[0119]

[0120]

[0121] From the above, it can be seen that, firstly, these optimization algorithms can automatically search in a wide range of hyperparameter spaces to find better hyperparameter combinations, thereby improving the efficiency and accuracy of hyperparameter tuning; compared with traditional manual tuning methods, these algorithms can greatly save time and labor costs, while avoiding the interference of human factors, making the tuning results more objective and reliable; secondly, these optimization algorithms have strong versatility and scalability; the method can be applied to various machine learning models and data sets, and is not limited by model type and scale; at the same time, with the continuous development of machine learning technology, new optimization algorithms and strategies can also be easily integrated into the system to meet changing needs; finally, by adopting these optimization algorithms, the invention also improves the performance of machine learning models; the optimized hyperparameters can enable the model to better fit the data during training, and improve the prediction accuracy and generalization ability of the model; this is undoubtedly an important improvement for machine learning models in practical application scenarios.

[0122] In summary, this invention optimizes the hyperparameters of the machine learning model based on random search, simulated annealing, particle swarm optimization, and genetic algorithm, and has achieved remarkable results, providing strong support for the development and application of machine learning technology.

[0123] A general black-box optimization method for hyperparameters in machine learning, such as Figure 3 As shown, the following steps are included:

[0124] S1. Collect training data and test data for machine learning models;

[0125] S2. Optimize the hyperparameters of the machine learning model;

[0126] S3. Use the optimized hyperparameters to train the machine learning model.

[0127] S4. Evaluate the trained machine learning model;

[0128] S5. Output the optimized hyperparameters and evaluation results.

[0129] In summary, the present invention has the following beneficial effects:

[0130] 1) Improve the efficiency and accuracy of hyperparameter tuning: The present invention adopts automated hyperparameter optimization methods, such as random search, simulated annealing, particle swarm optimization, genetic algorithm, etc., which can quickly find the optimal hyperparameter combination, significantly improving the efficiency and accuracy of hyperparameter tuning; compared with traditional manual tuning methods, this method greatly saves time and labor costs, while avoiding the interference of human factors, making the tuning results more objective and reliable;

[0131] 2) Universality and scalability: The system and method of the present invention are applicable to various machine learning models and data sets, and have strong universality and scalability. Regardless of the model type and size, the system can be used for hyperparameter optimization. At the same time, with the continuous development of machine learning technology, new optimization algorithms and strategies can also be easily integrated into the system to meet changing needs.

[0132] 3) Ease of use: The present invention lowers the threshold for machine learning parameter adjustment, allowing non-expert users to perform efficient hyperparameter optimization. Users do not need to pay attention to the specific algorithm implementation and context of the hyperparameter server, but only need to abide by certain conventions to perform general hyperparameter optimization work.

[0133] 4) Fast adaptation: When users introduce new algorithms, the system can quickly adapt and optimize without a lot of extra work;

[0134] 5) High availability and high scalability: This system supports large-scale applications. It only needs to expand the training nodes on the hyperparameter server side, and no adjustments are required on the user side. This high availability and high scalability enable the system to cope with various complex application scenarios.

[0135] 6) Built-in advanced algorithms: The system has built-in advanced hyperparameter optimization algorithms, such as Bayesian optimization, Gaussian process, grid search, etc. These algorithms can be directly put into production without additional development and debugging;

[0136] 7) Cross-platform optimization: This system has a wider range of applications and is suitable for a variety of machine learning platforms, providing users with more choices and flexibility.

[0137] The embodiment of the present application provides an electronic device applicable to the above-mentioned general machine learning hyperparameter black box optimization system, including:

[0138] Memory, used to protect computer programs and data;

[0139] Processor, used to run system programs.

[0140] An embodiment of the present application provides a computer storage medium, which is applicable to the above-mentioned general machine learning hyper-parameter black box optimization system, and performs hierarchical confidentiality management on the above-mentioned system and data in accordance with confidentiality management requirements.

[0141] Those skilled in the art will appreciate that the embodiments of the present application may be provided as a system or a computer program product. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0142] The present application is described with reference to the flowcharts and / or block diagrams of the devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0143] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0144] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0145] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0146] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0147] Computer readable media include permanent and non-permanent, removable and non-removable media, and can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include transitory media such as modulated data signals and carrier waves.

[0148] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, commodity or device including the elements.

[0149] The embodiments of the present invention are provided for the purpose of illustration and description. Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations of the present invention. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A general machine learning hyperparameter black box optimization system, characterized by: include: Hyperparameter optimization client, used to submit optimization requests; Hyperparameter optimization service system, used to process received client requests; Optimization algorithm configuration component, used to configure the hyperparameter optimization algorithm; Automatic stopping component, used to stop meaningless training sets; Hyperparameter optimization control panel, used to manage configuration, scheduling monitoring, user management, permission management, and system settings.

2. A general machine learning hyperparameter black box optimization system as claimed in claim 1, characterized in that: The hyper-parameter optimization client also includes a function for users to encapsulate requests that need to be processed, and the encapsulated information includes machine learning metadata and related resources.

3. A general machine learning hyperparameter black box optimization system as claimed in claim 1, characterized in that: The hyper-parameter optimization service system comprises: Data collection module, used to collect training data and test data for machine learning models; Hyperparameter optimization module, used to optimize the hyperparameters of machine learning models; Model training module, used to train machine learning models using optimized hyperparameters; Model evaluation module, used to evaluate the trained machine learning model; The result output module is used to output the optimized hyperparameters and evaluation results.

4. A general machine learning hyperparameter black box optimization system as claimed in claim 3, characterized in that: The hyperparameter optimization module optimizes the hyperparameters of the machine learning model based on random search, simulated annealing, particle swarm optimization, and genetic algorithm.

5. A general machine learning hyperparameter black box optimization system as claimed in claim 3, characterized in that: The model evaluation module uses accuracy, recall, and F1 value indicators to evaluate the machine learning model.

6. A general machine learning hyperparameter black box optimization system as claimed in claim 3, characterized in that: The result output module outputs the optimized hyperparameters and evaluation results in the form of tables and charts.

7. A general machine learning hyperparameter black box optimization system as claimed in claim 1, characterized in that: The optimization algorithm configuration component includes two types: built-in optimization algorithms and user-defined optimization algorithms, where the built-in algorithms include Bayesian optimization, Gaussian process, grid search, and random search.

8. A general machine learning hyperparameter black box optimization method as described in any one of claims 1 to 7, characterized in that: The following steps are involved: S1. Collect training data and test data for machine learning models; S2. Optimize the hyperparameters of the machine learning model; S3. Use the optimized hyperparameters to train the machine learning model. S4. Evaluate the trained machine learning model; S5. Output the optimized hyperparameters and evaluation results.

9. A processor, characterized in that: A general machine learning hyperparameter black-box optimization system configured to perform according to any one of claims 1 to 7.

10. 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 general machine learning hyperparameter black box optimization system described in any one of claims 1 to 7 is implemented.

Images