Model training method, device, electronic device and readable storage medium

Abstract

The present disclosure provides a model training method, device, electronic device and readable storage medium in the field of artificial intelligence technology, especially in the field of machine learning and text generation. The specific implementation scheme is: input the predicted sample into the encoding end of the model to be trained to obtain the feature vector of the predicted sample. According to the feature vector and the real result corresponding to the predicted sample, the real result is input into the decoding end to obtain the predicted result. According to the real result prediction result and the first loss function, a first loss value is calculated. The parameters of the model to be trained are adjusted according to the first loss value. Because the real result and the predicted result are taken into account when calculating the loss value, rather than simply relying on the real result. The model trained according to the loss value can refer to the real result and the predicted result at the same time, retain more diverse prediction results, and improve the robustness of the model and the diversity of the generated results.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, in particular to the fields of machine learning and text generation, and provides a model training method, device, electronic device and readable storage medium. Background Art

[0002] Generative models are widely used in text generation, summary generation, translation and other scenarios. When training a generative model, the teacher forcing method can be used to train the generative model. Among them, the cross entropy loss function can be used as the loss function for model training. The cross entropy loss is obtained by predicting the ground truth as the input of the decoder according to the cross entropy loss function. Summary of the invention

[0003] The present disclosure provides a model training method, device, electronic device and readable storage medium.

[0004] According to a first aspect of the present disclosure, a model training method is provided, comprising: inputting a prediction sample into an encoding end of a model to be trained to obtain a feature vector of the prediction sample. According to the feature vector and the real result corresponding to the prediction sample, inputting the real result into a decoding end to obtain a prediction result. According to the real result prediction result and a first loss function, a first loss value is calculated. According to the first loss value, the parameters of the model to be trained are adjusted.

[0005] According to a second aspect of the present disclosure, a model training device is provided, comprising: an input module, used to input a prediction sample into an encoding end of a model to be trained to obtain a feature vector of the prediction sample; a prediction module, used to input a real result into a decoding end according to the feature vector and the real result corresponding to the prediction sample to obtain a prediction result; a calculation module, used to calculate a first loss value according to the real result prediction result and a first loss function; and a training module, used to adjust the parameters of the model to be trained according to the first loss value.

[0006] According to a third aspect of the present disclosure, an electronic device is provided, comprising: at least one processor; and a memory connected to the at least one processor in communication. The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the method provided in the first aspect.

[0007] According to a fourth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, where the computer instructions are used to cause a computer to execute the method provided in the first aspect.

[0008] According to a fifth aspect of the present disclosure, a computer program product is provided, comprising a computer program, and when the computer program is executed by a processor, the method provided in the first aspect is implemented.

[0009] In the present disclosure, the actual results and predicted results are considered when calculating the loss value, rather than simply relying on the actual results. This allows the model trained based on the loss value to refer to both the actual results and the predicted results, retaining more diverse predicted results, thereby improving the robustness of the model and the diversity of the generated results.

[0010] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0012] Figure 1 is a flow chart of a model training method provided in the present disclosure;

[0013] Figure 2 is a flowchart of implementing S130 in a model training method provided in the present disclosure;

[0014] Figure 3 A structural block diagram of a model training device provided in an embodiment of the present disclosure;

[0015] Figure 4 A schematic block diagram of an example electronic device 400 that may be used to implement embodiments of the present disclosure is shown. DETAILED DESCRIPTION

[0016] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0017] Generative tasks are generally used in scenarios such as text generation, summary generation, translation, and machine dialogue. It is often necessary to train a corresponding generative model for each generative task. Since the input length and output length of the generative task are uncertain, the generative models that can be used include sequence-to-sequence (seq2seq) models, encoder-decoder models, etc. The basic principle is that the encoder (also called the encoding end) is used to analyze the input sequence, and the decoder (also called the decoding end) is used to generate the output sequence.

[0018] At present, most methods for training generative models use the teacher forcing method for training. The teacher forcing method means that at each prediction step, the real result is used as the input of the decoder. The decoder can learn to predict faster and the training process can converge faster.

[0019] However, the teacher forcing method will make the model overly dependent on real results, resulting in better results during training but worse results in actual application. For example, the synonyms generated by the prediction may be corrected to real results, causing the model's prediction results to lose diversity, that is, there is an exposure bias problem.

[0020] To this end, the present disclosure provides a model training method, device, electronic device and readable storage medium, which considers the real results and predicted results when calculating the loss value, rather than simply relying on the real results. The model trained according to the loss value can refer to the real results and predicted results at the same time, retain more diverse prediction results, and improve the robustness of the model and the diversity of the generated results.

[0021] In some embodiments, the model training method provided in the present disclosure can be applied to electronic devices such as personal computers, laptops, smart phones, tablet computers, servers, cloud servers, etc. with computing functions. The operating systems of these devices can be Android systems, Windows systems, iOS systems, Mac operating systems, or Harmony OS, Linux systems, etc., and the present disclosure does not limit this.

[0022] The model training method provided in the present disclosure can be applied to the training of models for various purposes. In the present disclosure, it is explained by taking the use for generating a model as an example.

[0023] Figure 1 It is a flow chart of a model training method provided in the present disclosure, which can be applied to any of the above-mentioned electronic devices.

[0024] refer to Figure 1, model training methods include:

[0025] S110, input the predicted sample into the encoding end of the model to be trained to obtain a feature vector of the predicted sample.

[0026] In some implementations, the model to be trained may be a generative model. In a generative model, each prediction sample may be a piece of text, such as "Hello", "How is the weather today", or "How are you", etc. The text may contain multiple languages, numbers, symbols, etc. The true result is a result pre-set according to the prediction sample. Each prediction sample corresponds to a true result. The true result of the prediction sample may be set differently depending on the purpose of the generative model. For example, when the generative model is used for machine dialogue, the true result of the prediction sample "Hello" may be "How can I help you?", and when the generative model is used for translation, the true result of the prediction sample "Hello" may be "Hello".

[0027] In some implementations, taking the seq2seq model as an example, the encoding end of the generation model may include multiple recurrent neural networks (RNNs), each of which may be a hidden state. When the prediction sample is input into the encoding end of the model to be trained, the first character of the text in the prediction sample is input into a hidden state, and then the output of the hidden state and the second character of the text are input into the next hidden state, until all characters are input into the hidden state, and the output of the last hidden state is used as the feature vector of the prediction sample, that is, the semantic vector of the text in the prediction sample.

[0028] S120. According to the feature vector and the actual result corresponding to the predicted sample, the actual result is input into the decoding end to obtain the predicted result.

[0029] In some implementations, the decoding end of the model to be trained may also include multiple RNNs, each of which may be a hidden state. In the generative model, the feature vector may be first input into the decoding end. If the output prediction result is different from the actual result after the feature vector is input into the first hidden state of the encoding end, the prediction result is discarded and the actual result is used as the input of the next hidden state. Until the prediction result is output. The prediction result may deviate from the actual result. For example, when the generative model is used for machine dialogue, the actual result of the prediction sample "Hello" may be "How can I help you?" The prediction result may be an incoherent sentence, such as "I can provide help"; or it may be a synonymous but different expression, such as "What can I do to help you?"

[0030] In some implementations, the prediction result can be determined from multiple candidate prediction results. For example, when making a prediction, multiple candidate prediction results with similar semantics may be obtained, and each candidate prediction result corresponds to a confidence level, that is, the probability that it is a true result. The prediction result can be determined from multiple candidate prediction results whose probabilities are greater than a preset probability threshold. The determination method includes but is not limited to: selecting the one with the highest confidence level, randomly selecting one of them, etc.

[0031] S130. Calculate a first loss value according to the actual result, the predicted result and the first loss function.

[0032] S140. Adjust the parameters of the model to be trained according to the first loss value.

[0033] In some embodiments, the first loss function may be a cross entropy loss function. Cross entropy can measure the degree of difference between two different probability distributions in the same random variable, that is, it is expressed as the difference between the distribution of the true prediction result and the distribution of the prediction result. The smaller the value of the cross entropy, the better the model prediction effect. The first loss value is the value calculated after the true result and the prediction result are input into the first loss function. Then, the training weights and bias terms can be updated by solving the update gradient according to the first loss value.

[0034] In this embodiment, by considering the actual results and the predicted results when calculating the loss value, rather than simply relying on the actual results, the model trained according to the loss value can refer to the actual results and the predicted results at the same time, retaining more diverse predicted results, thereby improving the robustness of the model and the diversity of the generated results.

[0035] Figure 2 It is a flowchart of implementing S130 in a model training method provided in the present disclosure.

[0036] In some implementations, the first loss function (Loss) may include a second loss function and a third loss function. Figure 2 To implement S130, you can follow these steps:

[0037] S1301. Input the true result into the second loss function to obtain a second loss value.

[0038] In some implementations, the second loss function is used to calculate the cross entropy loss value of the true result. For example, the second loss function can be Among them, y t For real results.

[0039] S1302: Input the prediction result into a third loss function to obtain a third loss value.

[0040] In some implementations, the third loss function is used to calculate the cross entropy loss value of the prediction result. For example, the third loss function can be Among them, y′ t is the prediction result. Among them, y′ t = argmax zEV logp(z|y<t,x), V is the set of all candidate prediction results.

[0041] S1303. Calculate the first loss value according to the second loss value, the third loss value, the weight of the second loss function, and the weight of the third loss function.

[0042] In some implementations, the second loss value can be multiplied by the weight of the second loss function to obtain a first loss parameter, the third loss value can be multiplied by the weight of the third loss function to obtain a second loss parameter, and then the first loss parameter and the second loss parameter are added and the inverse is taken to obtain the first loss value.

[0043] In some implementations, the weight of the second loss function is 1 minus the weight of the third loss function. When the probability value of the obtained prediction result meets the preset first condition, the weight of the third loss function is the preset weight coefficient multiplied by the trained step length, divided by the preset training step length. When the probability value of the obtained prediction result does not meet the preset first condition, the weight of the third loss function is 0.

[0044] The first loss function (Loss) can be expressed by the following formula:

[0045]

[0046] Among them, (1-α i ) is the weight of the second loss function, α i is the weight of the third loss function. i It can be expressed by the following formula:

[0047]

[0048] In some implementations, β can be set to 0.5 to ensure that α i The maximum value is no more than 0.5, step is the number of steps in the current training, step all is the total number of training steps, is the first preset condition, indicating that the sum of the probabilities of the top m candidate prediction results with the highest probabilities is greater than the number of samples γ, where γ is the probability threshold, N is the total number of samples, and θ is the threshold.

[0049] In some embodiments, when α i Not Compliant , it means that the training of the generative model is not sufficient and the confidence of the prediction result is not high. In this case, it is necessary to continue training according to the second loss function, so α i is 0, that is, the third loss function does not intervene in training for the time being.

[0050] When α i conform to , it means that the generative model has been trained sufficiently and can predict the true result or the synonym of the true result with a high probability. In this case, in order to make the trained generative model have more diverse output results, the third loss function can be introduced, that is, α i for When β=0.5, α i The value after running the total number of training steps is 0.5. That is, after the third loss function is introduced, the longer the training time, the higher the weight of the third loss function, that is, the stronger the intervention in the training. At the same time, the weight of the second loss function is (1-α i ), that is, when the third loss function is introduced, the intervention of the second loss function on training gradually weakens.

[0051] In this embodiment, by mixing the second loss function and the third loss function, the prediction results can be used to train the model, alleviate the overcorrection problem of the teacher forcing mechanism, and reduce exposure bias. In other words, no matter whether the input at the decoding end is the real result or the feature vector of the prediction sample, the probability that the output prediction result is consistent is greatly improved.

[0052] After testing, the generative model trained using the model training method provided by the present disclosure has a 30% improvement in diversity and basically the same accuracy compared to the generative model trained using only the teacher forcing method.

[0053] In this embodiment, by setting weights for the second loss function and the third loss function respectively, after meeting the preset first condition, the weight of the second loss function gradually decreases, and the third loss function is connected and gradually increases, so as to prevent the model from correcting the predicted results to the real results too early, so that the model trained according to the first loss function can take into account both the real results and the predicted results, with better robustness and better diversity of generated results.

[0054] It should be understood that the order of execution of each step in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present disclosure.

[0055] Corresponding to the model training method provided in the above embodiment, Figure 3This is a structural block diagram of a model training device provided in an embodiment of the present disclosure. For the sake of convenience, only the parts related to the embodiment of the present disclosure are shown.

[0056] Reference Figure 3 , a model training device, comprising:

[0057] The input module 21 is used to input the prediction sample into the encoding end of the model to be trained to obtain the feature vector of the prediction sample.

[0058] The prediction module 22 is used to input the real result into the decoding end according to the feature vector and the real result corresponding to the prediction sample to obtain the prediction result.

[0059] The calculation module 23 is used to calculate the first loss value according to the actual result prediction result and the first loss function.

[0060] The training module 24 is used to adjust the parameters of the model to be trained according to the first loss value.

[0061] In some implementations, the first loss function includes a second loss function and a third loss function;

[0062] The calculation module 23 is specifically used to input the real result into the second loss function to obtain the second loss value. The predicted result is input into the third loss function to obtain the third loss value. The first loss value is calculated according to the second loss value, the third loss value, the weight of the second loss function, and the weight of the third loss function.

[0063] In some implementations, the calculation module 23 is specifically configured to multiply the second loss value by the weight of the second loss function to obtain the first loss parameter. Multiply the third loss value by the weight of the third loss function to obtain the second loss parameter. Add the first loss parameter and the second loss parameter and take the opposite number to obtain the first loss value.

[0064] In some implementations, each prediction result corresponds to a probability value. When the probability value of the prediction result obtained meets the preset first condition, the weight of the third loss function is the preset weight coefficient multiplied by the trained step length, divided by the preset training step length. When the probability value of the prediction result obtained does not meet the preset first condition, the weight of the third loss function is 0.

[0065] In some implementations, the weight of the second loss function is 1 minus the weight of the third loss function.

[0066] In the technical solution disclosed herein, the acquisition, storage and application of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0067] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0068] Figure 4 A schematic block diagram of an example electronic device 400 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0069] like Figure 4 As shown, the device 400 includes a computing unit 401, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 402 or a computer program loaded from a storage unit 404 into a random access memory (RAM) 403. In the RAM 403, various programs and data required for the operation of the device 400 can also be stored. The computing unit 401, the ROM 402, and the RAM 403 are connected to each other via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0070] A number of components in the device 400 are connected to the I / O interface 405, including: an input unit 406, such as a keyboard, a mouse, etc.; an output unit 407, such as various types of displays, speakers, etc.; a storage unit 404, such as a disk, an optical disk, etc.; and a communication unit 409, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 409 allows the device 400 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0071] The computing unit 401 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 401 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 401 performs the various methods and processes described above, such as the model training method. For example, in some embodiments, the model training method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 404. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 400 via ROM 402 and / or communication unit 409. When the computer program is loaded into RAM 403 and executed by the computing unit 401, one or more steps of the model training method described above may be performed. Alternatively, in other embodiments, the computing unit 401 may be configured to perform the model training method in any other appropriate manner (e.g., by means of firmware).

[0072] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0073] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0074] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0075] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0076] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0077] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0078] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0079] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A model training method, characterized in that: include: Inputting a prediction sample into the encoding end of the model to be trained to obtain a feature vector of the prediction sample, wherein the prediction sample is text; According to the feature vector and the real result corresponding to the predicted sample, the real result is input into a decoding end to obtain a predicted result; Inputting the true result into a second loss function to obtain a second loss value, wherein the second loss function is used to calculate the cross entropy loss value of the true result; Inputting the prediction result into a third loss function to obtain a third loss value, wherein the third loss function is used to calculate the cross entropy loss value of the prediction result; Calculate a first loss value according to the second loss value, the third loss value, the weight of the second loss function, and the weight of the third loss function; Adjust the parameters of the model to be trained according to the first loss value.

2. The method according to claim 1, characterized in that The calculating the first loss value according to the second loss value, the third loss value, the weight of the second loss function, and the weight of the third loss function includes: Multiplying the second loss value by the weight of the second loss function to obtain a first loss parameter; Multiplying the third loss value by the weight of the third loss function to obtain a second loss parameter; The first loss parameter and the second loss parameter are added and the inverse thereof is taken to obtain the first loss value.

3. The method according to claim 1 or 2, characterized in that: Each of the prediction results corresponds to a probability value, and the method further comprises: When the obtained probability value of the prediction result meets the preset first condition, the weight of the third loss function is the preset weight coefficient multiplied by the trained step length divided by the preset training step length; When the obtained probability value of the prediction result does not meet the preset first condition, the weight of the third loss function is 0.

4. The method according to claim 3, characterized in that: The method also includes: the weight of the second loss function is 1 minus the weight of the third loss function.

5. A model training device, characterized in that: include: An input module, used to input a prediction sample into the encoding end of the model to be trained to obtain a feature vector of the prediction sample, wherein the prediction sample is text; A prediction module, configured to input the real result into a decoding end according to the feature vector and the real result corresponding to the prediction sample, so as to obtain a prediction result; A calculation module, used to input the real result into a second loss function to obtain a second loss value; input the predicted result into a third loss function to obtain a third loss value; calculate the first loss value according to the second loss value, the third loss value, the weight of the second loss function, and the weight of the third loss function, the second loss function is used to calculate the cross entropy loss value of the real result, and the third loss function is used to calculate the cross entropy loss value of the predicted result; A training module is used to adjust the parameters of the model to be trained according to the first loss value.

6. The device according to claim 5, characterized in that The calculation module is specifically configured to multiply the second loss value by the weight of the second loss function to obtain a first loss parameter; Multiplying the third loss value by the weight of the third loss function to obtain a second loss parameter; The first loss parameter and the second loss parameter are added and the inverse thereof is taken to obtain the first loss value.

7. The device according to claim 5 or 6, characterized in that Each of the prediction results corresponds to a probability value; When the obtained probability value of the prediction result meets the preset first condition, the weight of the third loss function is the preset weight coefficient multiplied by the trained step length divided by the preset training step length; When the obtained probability value of the prediction result does not meet the preset first condition, the weight of the third loss function is 0.

8. The device according to claim 7, characterized in that The weight of the second loss function is 1 minus the weight of the third loss function.

9. An electronic device, comprising: at least one processor; and a memory communicatively coupled to the at least one processor; It is characterized in that the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method according to any one of claims 1 to 4.

10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-4.

11. A computer program product, comprising a computer program, characterized in that The computer program implements the method according to any one of claims 1 to 4 when executed by a processor.

Images