Method, device, and electronic device for processing user input data by combining RPA and AI

Abstract

The method, device, and electronic device for processing user input data in combination with RPA and AI provided in this application are applied to natural language processing (NLP) of artificial intelligence AI, by obtaining user input data to be processed; processing the user input data to obtain attribute probability distribution and matching probability distribution; wherein the attribute probability distribution is used to represent the degree of similarity between the user input data and each attribute value under any attribute type, and the matching probability distribution is used to represent the degree of matching between the user input data and each candidate result; the information state of the user input data composed of the attribute probability distribution and the matching probability distribution is input into a preset network model and the dialogue data is output for speech understanding, so that the input data including noise can be directly processed, effectively improving the recognition accuracy of the input data, and at the same time, the user does not need to input data with the same semantic content multiple times, simplifying the user's operation process.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of data processing, and in particular, to a method, device, and electronic device for processing user input data by combining RPA (Robotic Process Automation) and AI (Artificial Intelligence). Background Art

[0002] Robotic Process Automation (RPA) uses specialized "robot software" to simulate human operations on computers, automatically executing process tasks according to rules. Artificial Intelligence (AI) is the technical science that studies and develops theories, methods, technologies, and application systems for simulating, extending, and expanding human intelligence. In the process of recognizing the semantics of natural language, the presence of noise in the language information can seriously affect the accuracy of the results.

[0003] In the field of computer science, a common approach to improving accuracy in natural language processing (NLP) applications is to collect noisy speech information multiple times. For example, this involves receiving multiple user inputs of the same semantic meaning, such as speech data or sentences. Each input is then subjected to the same recognition processing for speech understanding until an accurate recognition result is obtained.

[0004] However, this method requires the user to input data repeatedly, which is cumbersome and has low recognition processing efficiency. It also fails to improve the recognition accuracy of input data with noise. Summary of the Invention

[0005] To address the above issues, the present application provides a method, device, electronic device, and storage medium for processing user input data by combining RPA and AI.

[0006] In a first aspect, this application provides a method for processing user input data by combining RPA and AI, which is applied to natural language processing (NLP) of artificial intelligence AI, including:

[0007] Get the user input data to be processed;

[0008] Processing the user input data to obtain an attribute probability distribution and a matching probability distribution; wherein the attribute probability distribution is used to represent the similarity between the user input data and each attribute value under any attribute type, and the matching probability distribution is used to represent the matching degree between the user input data and each candidate result;

[0009] The information state of the user input data composed of the attribute probability distribution and the matching probability distribution is input into a preset network model and the dialogue data is output for speech understanding.

[0010] In some embodiments of the present application, after obtaining the user input data to be processed, the method further includes:

[0011] Determine whether this is the first round of processing of the user input data to be processed;

[0012] If so, the processing method further includes:

[0013] According to the user input data, the results stored in the preset result candidate library are screened to obtain a number of candidate results;

[0014] If not, the user input data of this round processes the candidate results, attribute probability distribution and matching probability distribution of the previous round to obtain the information state of the user input data composed of the attribute probability distribution of this round and the matching probability distribution, so as to input it into the preset network model and output the conversation data.

[0015] In some embodiments of the present application, after outputting the conversation data, the method further includes:

[0016] The processing round is updated, and according to the relationship between the updated processing round value and the preset number threshold, it is determined whether to return to the step of obtaining the user input data to be processed.

[0017] In some embodiments of the present application, the attribute probability distribution includes an attribute probability value and a probability confidence level;

[0018] Processing the user input data to obtain attribute probability distribution includes:

[0019] Determining the attribute probability value according to the similarity between the user input data and the attribute value;

[0020] The probability confidence corresponding to the attribute probability value is determined based on historical user input data.

[0021] In some embodiments of the present application, each candidate result is stored in a preset candidate result library, and the preset candidate result library also stores the attribute type of each candidate result and the corresponding attribute value;

[0022] The user input data is processed to obtain a matching probability distribution, including:

[0023] Determine the degree of association between the user input data and each candidate result based on the attribute probability distribution of the user input data and each attribute value under any attribute type, and the attribute value of the attribute type of each candidate result, and obtain a matching probability between the user input data and each candidate result;

[0024] The matching probability of each candidate result constitutes the matching probability distribution.

[0025] In some embodiments of the present application, further comprising:

[0026] Establish the network model to be trained;

[0027] Based on a reinforcement learning algorithm, network parameters in the network model to be trained are trained using the data samples in the training sample set, so that the trained network model outputs sample conversation data of the data samples in the training sample set according to the sample information state of the data samples in the training sample set;

[0028] The trained network model is used as the preset network model.

[0029] In some embodiments of the present application, the training samples include data samples from multiple rounds of training;

[0030] The step of establishing a network model to be trained includes:

[0031] Establish the initial network parameters and initial reward function of the network model to be trained;

[0032] The method of training the network parameters of the network model to be trained using the data samples in the training sample set based on the reinforcement learning algorithm includes:

[0033] According to the reinforcement learning algorithm, the initial network parameters of the network model to be trained are trained using the data samples of any round of training to obtain sample conversation data;

[0034] Updating the initial reward function according to the sample conversation data to obtain target network parameters and target reward function of the network model to be trained;

[0035] Determine whether the training result is the expected result, and if so, use the trained network model as the preset network model;

[0036] If not, according to the reinforcement learning algorithm, the target network parameters in the network model to be trained are trained using the data samples of the next round of training to obtain sample conversation data, and the step of updating the initial reward function according to the sample conversation data is repeated until the trained network model is obtained.

[0037] In a second aspect, the present application provides a device for processing user input data by combining RPA and AI, which is applied to natural language processing (NLP) of artificial intelligence (AI), including:

[0038] An acquisition module is used to obtain user input data to be processed;

[0039] a data processing module, configured to process the user input data to obtain an attribute probability distribution and a matching probability distribution; wherein the attribute probability distribution is used to indicate the degree of similarity between the user input data and each attribute value under any attribute type, and the matching probability distribution is used to indicate the degree of matching between the user input data and each candidate result;

[0040] The network model processing module is used to input the information state of the user input data composed of the attribute probability distribution and the matching probability distribution into a preset network model and output the dialogue data for speech understanding.

[0041] In a third aspect, the present application provides an electronic device, comprising: at least one processor and a memory;

[0042] The memory stores computer-executable instructions;

[0043] The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs the method as described in any of the preceding items.

[0044] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions. When a processor executes the computer-executable instructions, the method described in any of the preceding items is implemented.

[0045] The method, device, electronic device and storage medium for processing user input data in combination with RPA and AI provided in this application obtain user input data to be processed; process the user input data to obtain attribute probability distribution and matching probability distribution; wherein the attribute probability distribution is used to represent the degree of similarity between the user input data and each attribute value under any attribute type, and the matching probability distribution is used to represent the degree of matching between the user input data and each candidate result; the information state of the user input data composed of the attribute probability distribution and the matching probability distribution is input into a preset network model and the dialogue data is output for speech understanding, so that the input data including noise can be directly processed, effectively improving the recognition accuracy of the input data. At the same time, the user does not need to input data with the same semantic content multiple times, simplifying the user's operation process. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the examples of this application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the examples or prior art descriptions. Obviously, the drawings described below are some examples of this application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0047] Figure 1 A schematic diagram of the network architecture on which this application is based;

[0048] Figure 2 A flowchart of a method for processing user input data by combining RPA and AI, provided as an example of this application;

[0049] Figure 3 A flowchart of another method for processing user input data by combining RPA and AI provided as an example of this application;

[0050] Figure 4 A flowchart of another method for processing user input data by combining RPA and AI provided as an example of this application;

[0051] Figure 5 This is a schematic diagram of the structure of a device for processing user input data that combines RPA and AI, provided in this application;

[0052] Figure 6 Schematic diagram of the hardware structure of the electronic device provided for this application example. DETAILED DESCRIPTION

[0053] To make the purpose, technical solutions, and advantages of the examples of this application more clear, the technical solutions in the examples of this application will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the examples described are only part of the examples of this application, not all of them. Based on the examples in this application, all other examples obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of this application.

[0054] In the process of recognizing the semantics of natural language, the presence of noise in language information will seriously affect the accuracy of the results.

[0055] In the prior art, a common method for improving accuracy is to collect noisy speech information multiple times. For example, this involves receiving multiple user inputs of the same semantic meaning in speech or sentence information. Each piece of collected speech information is then subjected to the same recognition process until an accurate recognition result is obtained.

[0056] However, this method requires the user to repeatedly input language information, which has a cumbersome operation process, low recognition processing efficiency, and cannot improve the recognition accuracy of language information with noise.

[0057] To address the above issues, the present application provides a method, device, electronic device, and storage medium for processing user input data by combining RPA and AI.

[0058] It should be noted that the "processing" in this application refers to the data processing process that combines robotic process automation (RPA) and artificial intelligence (AI). That is to say, the data processing process is a fully automated data processing process, and the data processing process is also combined with artificial intelligence (AI) to realize the fully automated parsing and processing of user input data.

[0059] This application can be specifically applied to Natural Language Processing (NLP) in artificial intelligence AI, which is a field of computer science, artificial intelligence, and linguistics that focuses on the interaction between computers and human (natural) language.

[0060] For example, the present application realizes the full-process automatic identification of the acquired user input data based on the fully automated data processing process, executes the steps of automatically processing and parsing the user input data to obtain the attribute probability distribution and the matching probability distribution, and then combines some network models in the natural language processing NLP of artificial intelligence AI to process the attribute probability distribution and the matching probability distribution to obtain dialogue data for speech understanding.

[0061] refer to Figure 1 , Figure 1 This is a schematic diagram of the network architecture on which this application is based, such as Figure 1 As shown, a network architecture based on which the present application is based may include a device 2 for processing user input data in combination with RPA and AI, and a client 1.

[0062] Among them, the device 2 for processing user input data combining RPA and AI is hardware or software that can interact with the client 1 through the network, which can be used to execute the methods described in the following examples. The user can input the language information to be processed from the client 1, and the device 2 for processing user input data combining RPA and AI can capture the language information to be processed from the client 1 and output it to the device 2 for processing user input data combining RPA and AI.

[0063] When the device 2 for processing user input data that combines RPA and AI is hardware, it includes a cloud server with computing capabilities. When the device 2 for processing user input data that combines RPA and AI is software, it can be installed in an electronic device with computing capabilities, including but not limited to laptops and desktop computers.

[0064] Furthermore, client 1 refers to software or a platform capable of interacting with users. It can be loaded onto an electronic device in the form of a software interface or web interface for user operation. Through the network, client 1 and device 2, which processes user input data and combines RPA and AI, can communicate and exchange data.

[0065] First, reference Figure 2 , Figure 2 This is a flowchart of a method for processing user input data by combining RPA and AI provided in this application example. The method for processing user input data by combining RPA and AI provided in this application example includes:

[0066] Step 101: Obtain user input data to be processed.

[0067] It should be noted that the execution entity of the method for processing user input data by combining RPA and AI provided in the example of this application is a device for processing user input data by combining RPA and AI.

[0068] First, the processing device captures or receives data from the client to obtain the user input data to be processed. Specifically, this user input data can be voice information input by the user through audio or text information input by the user through a text input method. The content of the input data includes, but is not limited to, address navigation queries, encyclopedia questions, and intelligent question-and-answer services. The corresponding client types include, but are not limited to, intelligent navigation applications, encyclopedia queries, intelligent customer service platforms, and intelligent conversational robots.

[0069] In addition, after capturing or receiving data, the processing device may also preprocess the data. Generally, the preprocessing implementation method varies based on the form of the input data. For example, when the input data is voice data, the preprocessing will include voice-to-text conversion, thereby converting the voice data into computer-recognizable data to be processed; when the input data is text data, the preprocessing may include operations such as semantic segmentation and word segmentation on the text data, thereby converting the text data into computer-recognizable data to be processed.

[0070] Step 102: Process the user input data to obtain an attribute probability distribution and a matching probability distribution, wherein the attribute probability distribution is used to indicate the degree of similarity between the user input data and each attribute value under any attribute type, and the matching probability distribution is used to indicate the degree of matching between the user input data and each candidate result.

[0071] In order to accurately identify and process user input data, in this application, the attribute probability distribution and matching probability distribution corresponding to the user input data will be obtained respectively.

[0072] The attribute probability distribution can include an attribute probability value and a probability confidence level. The determination process is as follows: first, the attribute probability value is determined based on the similarity between the user input data and the attribute value; then, the probability confidence level corresponding to the attribute probability value is determined based on historical user input data.

[0073] It should be noted that the attribute type is used to represent the inherent attribute information of the data, which varies according to the different fields involved in the data. For example, if the scene is a navigation scene, and the input data is the location that needs to be navigated or queried, then the attribute type of the input data is the address attribute type "city", "district", "street", "business district", etc. Correspondingly, the attribute value refers to the attribute value under the attribute type. For example, the attribute value of the attribute type "city" may include "Beijing City", "Taipei City", etc. In the example of this application, each attribute type and its attribute value can be pre-stored in the processing device.

[0074] For the user input data "I want to go to Fuchengmen, Xicheng District", the similarity between the user input data and the attribute value can be determined. For example, if the attribute type "district" in the processing device stores multiple attribute values ​​such as "Xicheng District", "Yizhuang Development Zone", and "Dongcheng District"; based on the similarity comparison between the input data "Xicheng District" and the above attribute values ​​in pronunciation trends and the number of syllables, it can be seen that the attribute probability value of "Xicheng District" is the largest, "Dongcheng District" is second, and the attribute probability value of "Yizhuang Development Zone" is relatively low.

[0075] Subsequently, it is necessary to determine the probability confidence level corresponding to the attribute probability value based on the historical user input data. For example, in the historical user input data, the user once input "I want to go to Fuchengmen, Xicheng District", and the processing device determines that the user intended to enter "Fuchengmen (address) - Xicheng District, Beijing (city)", then after receiving the input data "I want to go to Fuchengmen, Xicheng District" this time, the processing device will believe that the user clearly entered "Xicheng District" into the processing device, rather than other addresses. In this case, the probability confidence level corresponding to the attribute value "Xicheng District" is higher; conversely, in the historical user input data, after the user entered "I want to go to Fuchengmen, Xicheng District", the processing device determines that the user actually intended to enter "Fuchengmen (address) - Taipei City (city) - Zhongzheng District (district)", in this case, although the attribute probability value of "Xicheng District" is higher, its probability confidence level is lower.

[0076] Through the above method, it can be determined that the user input data is based on different attribute types, and the attribute probability value and corresponding probability confidence of each attribute value under the attribute type can be obtained. All attribute probability values ​​and probability confidences will constitute the attribute probability distribution of the user input data.

[0077] In addition, the processing device also determines a matching probability distribution for the user input data. The matching probability distribution is used to represent the degree of match between the user input data and each candidate result. Each candidate result is stored in a preset candidate result library, which also stores the attribute type and corresponding attribute value of each candidate result. As previously mentioned, the attribute type refers to the inherent attribute information of the data, which varies depending on the field involved in the data, while the attribute value refers to the attribute value under the attribute type.

[0078] The attribute type of each candidate result and the corresponding attribute value can be expressed as follows:

[0079] [Fuchengmen (address): Beijing (city), Xicheng District (district), empty (street), empty (business district)]; another example is [Fuchengmen (address): Taipei City (city), Zhongzheng District (district), empty (street), tourist attraction (business district)].

[0080] Accordingly, when processing the user input data to obtain the matching probability distribution, the following method is adopted: based on the attribute probability distribution of the user input data and each attribute value under any attribute type, as well as the attribute value of the attribute type of each candidate result, the degree of association between the user input data and each candidate result is determined to obtain the matching probability between the user input data and each candidate result; the matching probability of each candidate result constitutes the matching probability distribution.

[0081] For example, for the input data "I want to go to Fuchengmen in Xicheng District", the candidate results may be [Fuchengmen (address): Beijing (city), Xicheng District (district), empty (street), empty (business district)]; another example is [Fuchengmen (address): Taipei (city), Zhongzheng District (district), empty (street), tourist attraction (business district)].

[0082] Among them, if the attribute probability value and probability confidence of the input data with "Xicheng District" are both high, then the matching probability of the input data on the information attribute of Xicheng District (district) is also high. By counting the attribute probability value and probability confidence of the attribute value under each attribute type, the degree of association between the input data and the candidate results can be determined and expressed as a matching probability. For example, the matching probability of "I want to go to Fuchengmen, Xicheng District" and [Fuchengmen (address): Beijing (city), Xicheng District (district), empty (street), empty (business district)] is [0.9 (address), 0.7 (city), 0.6 (district), 0 (street), 0 (business district)]; and the matching probability of [Fuchengmen (address): Taipei (city), Zhongzheng District (district), empty (street), tourist attractions (business district)] is [0.9 (address), 0.2 (city), 0.2 (district), 0 (street), 1 (business district)]. Among them, the matching probability value under each attribute type is calculated independently, and the candidate result, that is, the address item, has the corresponding matching probability value associated with the matching probability value corresponding to each attribute type.

[0083] The matching probabilities of the above-mentioned candidate results will constitute the matching probability distribution of the user input data.

[0084] Step 103: Input the information state of the user input data composed of the attribute probability distribution and the matching probability distribution into a preset network model and output the dialogue data for speech understanding.

[0085] Finally, the above-mentioned attribute probability distribution and matching probability distribution are spliced ​​together to form an information matrix for representing the information state of the user input data, and are input into the network model for processing and outputting results.

[0086] Specifically, the preset network model can be a deep learning model, specifically a fully connected network. This preset network model can be used to analyze the state of the input language information and output corresponding response information. For example, for the aforementioned user input data "I want to go to Fuchengmen in Xicheng District", the network model's output processing result may be "Is the address in Xicheng District?" or "Did you mean Fuchengmen?" The user can then further respond to the obtained processing result until a satisfactory result is obtained.

[0087] Through the above-mentioned processing method, the processing device can directly process the input data including noise, effectively improving the recognition accuracy of the input data. At the same time, the user does not need to input data with the same semantic content multiple times, simplifying the user's operation process.

[0088] Based on the above example, Figure 3 A flowchart of another method for processing user input data by combining RPA and AI is provided as an example of this application.

[0089] like Figure 3 As shown, the method for processing user input data by combining RPA and AI includes:

[0090] Step 201: Obtain user input data to be processed.

[0091] Step 202: Determine whether this is the first round of processing the user input data to be processed, and obtain candidate results based on the determination result.

[0092] Step 203: Process the user input data to obtain an attribute probability distribution and a matching probability distribution; wherein the attribute probability distribution is used to indicate the degree of similarity between the user input data and each attribute value under any attribute type, and the matching probability distribution is used to indicate the degree of matching between the user input data and each candidate result.

[0093] Step 204: Input the information state of the user input data composed of the attribute probability distribution and the matching probability distribution into a preset network model and output conversation data.

[0094] Step 205: Update the processing round and determine whether to return to the step of obtaining the user input data to be processed based on the relationship between the updated processing round value and the preset number threshold.

[0095] In this embodiment, step 201, step 203 and step 204 are similar to step 101, step 102 and step 103 in the above embodiment, and their specific implementation methods are not repeated in this embodiment.

[0096] Different from the above-mentioned embodiment, this embodiment further includes a step of determining the processing round of the user input data.

[0097] Specifically, the processing round represents the number of times the user input data is received and processed from the time the processing device is triggered until the processing is terminated.

[0098] For example, the following question-answering dialogue processing results can be obtained using the processing device:

[0099] Question and answer trigger:

[0100] User: "I want to go to Fuchengmen."; Processing device: "Excuse me, is this Xicheng District?" (Processing round is 1).

[0101] User: “Yes.”; System: “You are going to Fuchengmen Bar” (processing round is 2).

[0102] User: “Yes.”

[0103] In the above processing process, when obtaining the user input data to be processed, it is determined whether this is the first processing round for the user input data to be processed.

[0104] When the processing round of the user input data to be processed is the first, the results stored in the preset result candidate library may be screened according to the user input data to obtain a number of candidate results.

[0105] Specifically, the preset candidate result library stores a large amount of historical conversation data. To improve processing efficiency, it is necessary to filter the results from the preset candidate result library before processing the user input data to be processed, and use the filtered results as the candidate results. Furthermore, field filtering can be based on the ES algorithm, or other algorithms can be used, which are not limited in this application.

[0106] When the processing round of the user input data to be processed is not the first, it generally means a response to the dialogue data of the previous processing round. Therefore, when processing the results of this round, the candidate results, attribute probability distribution and matching probability distribution of the previous round can be processed based on the user input data of this round to obtain the information state of the user input data composed of the attribute probability distribution of this round and the matching probability distribution, so as to input it into the preset network model and output the dialogue data.

[0107] Furthermore, in this embodiment, to ensure processing efficiency, the processing device may limit the number of rounds, i.e., set a threshold number of rounds to ensure that the number of rounds processed for a question and answer session does not exceed the threshold number of rounds. Therefore, after each output of conversation data, the processing device will update the number of rounds to determine the relationship between the updated number of rounds and the threshold number of rounds.

[0108] When the round value is less than or equal to the number threshold, the processing device continues to respond to and process the user input data; when the round value is greater than the number threshold, the processing device may no longer respond to and process the user input data, but instead send a prompt message to the user such as "The information you entered cannot be queried" or "Query error" or "The address cannot be found".

[0109] Of course, in other optional embodiments, when the dialog data output by the processing device meets the needs of the user, the processing device may also stop responding to and processing the user input data. For example, in the aforementioned example, when the dialog data output at processing round 2 is used to give a positive response to the dialog data "Are you going to Fuchengmen?", the processing device will no longer process the dialog data and send it to other modules to provide other services to the user, such as sending the result to the navigation module so that the navigation module can provide navigation services to the user.

[0110] On the basis of the above-mentioned embodiment, this embodiment limits the processing rounds of the processing device, thereby effectively avoiding the problem of low processing efficiency caused by too many rounds during question and answer.

[0111] Based on the above example, Figure 4 The flowchart of another method for processing user input data by combining RPA and AI is provided as an example of this application. Figure 4 As shown, the method for processing user input data by combining RPA and AI includes:

[0112] Step 301: Establish and train a network model to be trained to obtain a preset network model.

[0113] Step 302: Obtain user input data to be processed;

[0114] Step 303: Process the user input data to obtain an attribute probability distribution and a matching probability distribution; wherein the attribute probability distribution is used to indicate the degree of similarity between the user input data and each attribute value under any attribute type, and the matching probability distribution is used to indicate the degree of matching between the user input data and each candidate result;

[0115] Step 304: Input the information state of the user input data composed of the attribute probability distribution and the matching probability distribution into a preset network model and output conversation data.

[0116] In this embodiment, step 302, step 303 and step 304 are similar to step 101, step 102 and step 103 in the above embodiment, and their specific implementation methods are not repeated in this embodiment.

[0117] The process of establishing and training a network model to be trained to obtain a preset network model may specifically include:

[0118] Step 3011: Establish a network model to be trained;

[0119] Step 3012: Based on a reinforcement learning algorithm, the network parameters of the network model to be trained are trained using the data samples in the training sample set, so that the trained network model outputs sample conversation data of the data samples in the training sample set according to the sample information state of the data samples in the training sample set;

[0120] Step 3013: Use the trained network model as the preset network model.

[0121] Specifically, the training samples include data samples from multiple rounds of training. Each round of training data samples includes at least training data and expected results. The data samples and expected results can be derived from a user simulator. For example, a random simulation algorithm can be used to simulate user demand information.

[0122] Subsequently, according to the reinforcement learning algorithm, the initial network parameters in the network model to be trained can be trained using the data samples of any round of training to obtain sample conversation data; the initial reward function can be updated according to the sample conversation data to obtain the target network parameters and target reward function of the network model to be trained.

[0123] Then, it is determined whether the training result is the expected result.

[0124] If the training result is as expected, the trained network model is used as the preset network model. If the training result is not as expected, the above process is repeated. That is, according to the reinforcement learning algorithm, the target network parameters of the network model to be trained are trained using the data samples of the next round of training to obtain sample conversation data, and the step of updating the initial reward function based on the sample conversation data is repeated until a trained network model is obtained.

[0125] Each data sample is processed similarly to the above method to obtain an information state consisting of an attribute probability distribution and a matching probability distribution, which is then input into the network model to be trained, causing the network model to output the training results, i.e., the conversation data. Furthermore, regarding the reward function, when the initial reward function is r_t, after completing one round of processing, if the question-answering task is completed, the reward function is set to 10, and if the question-answering task is not completed, the reward function r_t is reduced by 1. This encourages the model to complete the question-answering task with the fewest possible processing rounds, which is consistent with the expected results.

[0126] On the basis of the aforementioned embodiment, by using the reward function to train the network model, it is possible to effectively improve recognition efficiency, reduce processing rounds, and provide users with a better question-and-answer experience.

[0127] Secondly, Figure 5 This is a schematic diagram of the structure of a device for processing user input data that combines RPA and AI, which is applied to natural language processing (NLP) of artificial intelligence AI, such as Figure 5 As shown, the device for processing user input data by combining RPA and AI includes:

[0128] An acquisition module 10 is used to acquire user input data to be processed;

[0129] The data processing module 20 is configured to process the user input data to obtain an attribute probability distribution and a matching probability distribution; wherein the attribute probability distribution is used to indicate the degree of similarity between the user input data and each attribute value under any attribute type, and the matching probability distribution is used to indicate the degree of matching between the user input data and each candidate result;

[0130] The network model processing module 30 is used to input the information state of the user input data composed of the attribute probability distribution and the matching probability distribution into a preset network model and output the dialogue data for speech understanding.

[0131] In the optional example provided in the present application, the data processing module 20 is specifically configured to determine whether this is the first processing round for the user input data to be processed;

[0132] If so, the processing method further includes:

[0133] According to the user input data, the results stored in the preset result candidate library are screened to obtain a number of candidate results;

[0134] If not, the user input data of this round processes the candidate results, attribute probability distribution and matching probability distribution of the previous round to obtain the information state of the user input data composed of the attribute probability distribution of this round and the matching probability distribution, so as to input it into the preset network model and output the conversation data.

[0135] In the optional example provided in this application, the data processing module 20 is specifically configured to output the conversation data and further include: a turn determination module;

[0136] The round determination module is used to update the processing round and determine whether to return to the step of obtaining the user input data to be processed based on the relationship between the updated processing round value and a preset number threshold.

[0137] In the optional example provided in this application, the attribute probability distribution includes an attribute probability value and a probability confidence level;

[0138] The data processing module 20 is specifically used for:

[0139] Determining the attribute probability value according to the similarity between the user input data and the attribute value;

[0140] The probability confidence corresponding to the attribute probability value is determined based on historical user input data.

[0141] In the optional example provided in this application, each candidate result is stored in a preset candidate result library, and the preset candidate result library also stores the attribute type of each candidate result and the corresponding attribute value;

[0142] The user input data is processed to obtain a matching probability distribution, including: the data processing module 20 is specifically used to: determine the degree of association between the user input data and each candidate result based on the attribute probability distribution of the user input data and each attribute value under any attribute type, and the attribute value of the attribute type of each candidate result, to obtain the matching probability between the user input data and each candidate result; the matching probability of each candidate result constitutes the matching probability distribution.

[0143] In an optional example provided in this application, the processing device further includes: a training module;

[0144] The training module is used to establish a network model to be trained; based on a reinforcement learning algorithm, the network parameters in the network model to be trained are trained using data samples in a training sample set, so that the trained network model outputs sample conversation data of the data samples in the training sample set according to the sample information state of the data samples in the training sample set; and the trained network model is used as the preset network model.

[0145] In an optional example provided in this application, the training samples include data samples of multiple rounds of training;

[0146] The training module is used to establish the initial network parameters and initial reward function of the network model to be trained; it is also used to train the initial network parameters in the network model to be trained using the data samples of any round of training according to the reinforcement learning algorithm to obtain sample dialogue data; update the initial reward function according to the sample dialogue data to obtain the target network parameters and target reward function of the network model to be trained; judge whether the training result is the expected result, if so, use the trained network model as the preset network model; if not, use the data samples of the next round of training according to the reinforcement learning algorithm to train the target network parameters in the network model to be trained to obtain sample dialogue data, and repeat the step of updating the initial reward function according to the sample dialogue data until the trained network model is obtained.

[0147] The device for processing user input data provided by the present application combines RPA and AI, thereby obtaining user input data to be processed; processing the user input data to obtain an attribute probability distribution and a matching probability distribution; wherein the attribute probability distribution is used to represent the degree of similarity between the user input data and each attribute value under any attribute type, and the matching probability distribution is used to represent the degree of matching between the user input data and each candidate result; the information state of the user input data constituted by the attribute probability distribution and the matching probability distribution is input into a preset network model and dialogue data is output, thereby directly processing input data including noise, effectively improving the recognition accuracy of the input data, and at the same time, the user does not need to input data with the same semantic content multiple times, simplifying the user's operation process.

[0148] On the other hand, this example also provides an electronic device that can be used to execute the technical solution of the above method example. Its implementation principle and technical effects are similar, and this example will not be repeated here.

[0149] refer to Figure 6 It shows a schematic diagram of the structure of an electronic device 900 suitable for implementing the examples of the present application. The electronic device 900 can be a terminal device or a server. The terminal device can include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, personal digital assistants (PDAs), tablet computers (Portable Android Devices, PADs), portable multimedia players (PMPs), in-vehicle terminals (such as in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 6 The electronic device shown is only an example and should not limit the functions and scope of use of the examples of this application.

[0150] like Figure 6 As shown, the electronic device 900 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 901, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 902 or a program loaded from a storage device 908 into a random access memory (RAM) 903. Various programs and data required for the operation of the electronic device 900 are also stored in the RAM 903. The processing device 901, the ROM 902, and the RAM 903 are connected to each other via a bus 904. An input / output (I / O) interface 905 is also connected to the bus 904.

[0151] Typically, the following devices may be connected to the I / O interface 905: an input device 906 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 907 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 908 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 909. The communication device 909 may allow the electronic device 900 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 6 The electronic device 900 is shown with various devices, but it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed instead.

[0152] In particular, according to an example of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an example of the present application includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes program code for executing the method shown in the flowchart. In such an example, the computer program can be downloaded and installed from the network via the communication device 909, or installed from the storage device 908, or installed from the ROM 902. When the computer program is executed by the processing device 901, the above-mentioned functions defined in the method of the example of the present application are performed.

[0153] It should be noted that the computer-readable medium mentioned above in this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, 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 above. In this application, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, device, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. This propagated data signal can take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium may be transmitted using any suitable medium, including but not limited to wires, optical cables, RF (radio frequency), etc., or any suitable combination thereof.

[0154] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0155] The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device executes the method shown in the above example.

[0156] The computer program code for performing the operations of the present application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, using an Internet service provider to connect via the Internet).

[0157] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various examples of the present application. In this regard, each box in the flow chart or block diagram can represent a module, program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the prescribed logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a sequence different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a dedicated hardware-based system that performs the prescribed function or operation, or can be implemented by a combination of dedicated hardware and computer instructions.

[0158] The units described in the examples of this application may be implemented in software or hardware. In some cases, the name of a unit does not limit the unit itself. For example, the first acquisition unit may also be described as a "unit for acquiring at least two Internet Protocol addresses."

[0159] The functions described above herein may be performed, at least in part, by one or more hardware logic components. For example, and without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chip (SOCs), complex programmable logic devices (CPLDs), and the like.

[0160] In the context of the present application, a machine-readable medium can be a tangible medium that can contain or store a program for use by an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can 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 can 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.

[0161] Here are some examples of this application.

[0162] In a first aspect, according to one or more examples of the present application, a method for processing user input data by combining RPA and AI includes:

[0163] Get the user input data to be processed;

[0164] Processing the user input data to obtain an attribute probability distribution and a matching probability distribution; wherein the attribute probability distribution is used to represent the similarity between the user input data and each attribute value under any attribute type, and the matching probability distribution is used to represent the matching degree between the user input data and each candidate result;

[0165] The information state of the user input data composed of the attribute probability distribution and the matching probability distribution is input into a preset network model and the dialogue data is output for speech understanding.

[0166] In the optional example provided in this application, after obtaining the user input data to be processed, the method further includes:

[0167] Determine whether this is the first round of processing of the user input data to be processed;

[0168] If so, the processing method further includes:

[0169] According to the user input data, the results stored in the preset result candidate library are screened to obtain a number of candidate results;

[0170] If not, the user input data of this round processes the candidate results, attribute probability distribution and matching probability distribution of the previous round to obtain the information state of the user input data composed of the attribute probability distribution of this round and the matching probability distribution, so as to input it into the preset network model and output the conversation data.

[0171] In the optional example provided in this application, after outputting the conversation data, the following is further included:

[0172] The processing round is updated, and according to the relationship between the updated processing round value and the preset number threshold, it is determined whether to return to the step of obtaining the user input data to be processed.

[0173] In the optional example provided in this application, the attribute probability distribution includes an attribute probability value and a probability confidence level;

[0174] Processing the user input data to obtain attribute probability distribution includes:

[0175] Determining the attribute probability value according to the similarity between the user input data and the attribute value;

[0176] The probability confidence corresponding to the attribute probability value is determined based on historical user input data.

[0177] In the optional example provided in this application, each candidate result is stored in a preset candidate result library, and the preset candidate result library also stores the attribute type of each candidate result and the corresponding attribute value;

[0178] The user input data is processed to obtain a matching probability distribution, including:

[0179] Determine the degree of association between the user input data and each candidate result based on the attribute probability distribution of the user input data and each attribute value under any attribute type, and the attribute value of the attribute type of each candidate result, and obtain a matching probability between the user input data and each candidate result;

[0180] The matching probability of each candidate result constitutes the matching probability distribution.

[0181] In an optional example provided in this application, the processing method further includes:

[0182] Establish the network model to be trained;

[0183] Based on a reinforcement learning algorithm, network parameters in the network model to be trained are trained using the data samples in the training sample set, so that the trained network model outputs sample conversation data of the data samples in the training sample set according to the sample information state of the data samples in the training sample set;

[0184] The trained network model is used as the preset network model.

[0185] In an optional example provided in this application, the training samples include data samples of multiple rounds of training;

[0186] The step of establishing a network model to be trained includes:

[0187] Establish the initial network parameters and initial reward function of the network model to be trained;

[0188] The method of training the network parameters of the network model to be trained using the data samples in the training sample set based on the reinforcement learning algorithm includes:

[0189] According to the reinforcement learning algorithm, the initial network parameters of the network model to be trained are trained using the data samples of any round of training to obtain sample conversation data;

[0190] Updating the initial reward function according to the sample conversation data to obtain target network parameters and target reward function of the network model to be trained;

[0191] Determine whether the training result is the expected result, and if so, use the trained network model as the preset network model;

[0192] If not, according to the reinforcement learning algorithm, the target network parameters in the network model to be trained are trained using the data samples of the next round of training to obtain sample conversation data, and the step of updating the initial reward function according to the sample conversation data is repeated until the trained network model is obtained.

[0193] In a second aspect, according to one or more examples of the present application, a device for processing user input data that combines RPA and AI includes:

[0194] An acquisition module is used to obtain user input data to be processed;

[0195] a data processing module, configured to process the user input data to obtain an attribute probability distribution and a matching probability distribution; wherein the attribute probability distribution is used to indicate the degree of similarity between the user input data and each attribute value under any attribute type, and the matching probability distribution is used to indicate the degree of matching between the user input data and each candidate result;

[0196] The network model processing module is used to input the information state of the user input data composed of the attribute probability distribution and the matching probability distribution into a preset network model and output conversation data.

[0197] In the optional example provided in the present application, the data processing module is specifically used to determine whether this is the first processing round of the user input data to be processed;

[0198] If so, the processing method further includes:

[0199] According to the user input data, the results stored in the preset result candidate library are screened to obtain a number of candidate results;

[0200] If not, the user input data of this round processes the candidate results, attribute probability distribution and matching probability distribution of the previous round to obtain the information state of the user input data composed of the attribute probability distribution of this round and the matching probability distribution, so as to input it into the preset network model and output the conversation data.

[0201] In the optional example provided in the present application, the data processing module is specifically configured to output the conversation data and further include: a turn determination module;

[0202] The round determination module is used to update the processing round and determine whether to return to the step of obtaining the user input data to be processed based on the relationship between the updated processing round value and a preset number threshold.

[0203] In the optional example provided in this application, the attribute probability distribution includes an attribute probability value and a probability confidence level;

[0204] The data processing module is specifically used for:

[0205] Determining the attribute probability value according to the similarity between the user input data and the attribute value;

[0206] The probability confidence corresponding to the attribute probability value is determined based on historical user input data.

[0207] In the optional example provided in this application, each candidate result is stored in a preset candidate result library, and the preset candidate result library also stores the attribute type of each candidate result and the corresponding attribute value;

[0208] The user input data is processed to obtain a matching probability distribution, including: the data processing module is specifically used to: determine the degree of association between the user input data and each candidate result based on the attribute probability distribution of the user input data and each attribute value under any attribute type, and the attribute value of the attribute type of each candidate result, to obtain the matching probability between the user input data and each candidate result; the matching probability of each candidate result constitutes the matching probability distribution.

[0209] In an optional example provided in this application, the processing device further includes: a training module;

[0210] The training module is used to establish a network model to be trained; based on a reinforcement learning algorithm, the network parameters in the network model to be trained are trained using data samples in a training sample set, so that the trained network model outputs sample conversation data of the data samples in the training sample set according to the sample information state of the data samples in the training sample set; and the trained network model is used as the preset network model.

[0211] In an optional example provided in this application, the training samples include data samples of multiple rounds of training;

[0212] The training module is used to establish the initial network parameters and initial reward function of the network model to be trained; it is also used to train the initial network parameters in the network model to be trained using the data samples of any round of training according to the reinforcement learning algorithm to obtain sample dialogue data; update the initial reward function according to the sample dialogue data to obtain the target network parameters and target reward function of the network model to be trained; judge whether the training result is the expected result, if so, use the trained network model as the preset network model; if not, use the data samples of the next round of training according to the reinforcement learning algorithm to train the target network parameters in the network model to be trained to obtain sample dialogue data, and repeat the step of updating the initial reward function according to the sample dialogue data until the trained network model is obtained.

[0213] In a third aspect, according to one or more examples of the present application, an electronic device includes: at least one processor and a memory;

[0214] The memory stores computer-executable instructions;

[0215] The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs the method as described above.

[0216] In a fourth aspect, according to one or more examples of the present application, a computer-readable storage medium stores computer-executable instructions, and when a processor executes the computer-executable instructions, the method described above is implemented.

[0217] The above description is merely a preferred example of the present application and an illustration of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this application is not limited to the technical solutions formed by a specific combination of the above-mentioned technical features, but also encompasses other technical solutions formed by any combination of the above-mentioned technical features or their equivalents without departing from the above-mentioned disclosed concepts. For example, a technical solution formed by replacing the above-mentioned features with (but not limited to) technical features with similar functions disclosed in this application.

[0218] In addition, although each operation is described in a specific order, this should not be understood as requiring these operations to be performed in the specific order shown or in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details have been included in the above discussion, these should not be interpreted as limiting the scope of the application. Some features described in the context of a separate example can also be implemented in a single example in combination. On the contrary, the various features described in the context of a single example can also be implemented in multiple examples individually or in any suitable sub-combination.

[0219] Although the subject matter has been described in language specific to structural features and / or methodological logical acts, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are merely example forms of implementing the claims.

[0220] The purpose, technical solution, and advantages of the examples are more clearly understood. The following will provide a clear and complete description of the technical solutions in the examples of this application in conjunction with the accompanying drawings. Obviously, the examples described are only part of the examples of this application, not all of them. All other examples derived by persons of ordinary skill in the art based on the examples in this application without creative effort fall within the scope of protection of this application.

Claims

1. A method for processing user input data by combining RPA and AI, characterized in that: include: Get the user input data to be processed; Processing the user input data to obtain an attribute probability distribution and a matching probability distribution; wherein the attribute probability distribution is used to represent the similarity between the user input data and each attribute value under any attribute type, and the matching probability distribution is used to represent the matching degree between the user input data and each candidate result; Inputting the information state of the user input data composed of the attribute probability distribution and the matching probability distribution into a preset network model and outputting conversation data; After obtaining the user input data to be processed, the method further includes: Determine whether this is the first round of processing of the user input data to be processed; If so, the method includes: According to the user input data, the results stored in the preset result candidate library are screened to obtain a number of candidate results; If not, the method further comprises: Process the candidate results, attribute probability distribution, and matching probability distribution of the previous round based on the user input data of this round to obtain the information state of the user input data composed of the attribute probability distribution of this round and the matching probability distribution, input it into the preset network model, and output conversation data; Each candidate result is stored in a preset candidate result library, which also stores the attribute type of each candidate result and the corresponding attribute value; the user input data is processed to obtain a matching probability distribution, including: Based on the attribute probability distribution of the user input data and each attribute value under any attribute type, as well as the attribute value of the attribute type of each candidate result, the degree of association between the user input data and each candidate result is determined, and the matching probability between the user input data and each candidate result is obtained; the matching probability of each candidate result constitutes the matching probability distribution.

2. The method according to claim 1, characterized in that After outputting the conversation data, it also includes: The processing round is updated, and according to the relationship between the updated processing round value and the preset number threshold, it is determined whether to return to the step of obtaining the user input data to be processed.

3. The method according to claim 1, characterized in that The attribute probability distribution includes an attribute probability value and a probability confidence level. Processing the user input data to obtain the attribute probability distribution includes: Determining the attribute probability value according to the similarity between the user input data and the attribute value; The probability confidence corresponding to the attribute probability value is determined based on historical user input data.

4. The method according to any one of claims 1 to 3, characterized in that Also includes: Establish the network model to be trained; Based on a reinforcement learning algorithm, network parameters in the network model to be trained are trained using the data samples in the training sample set, so that the trained network model outputs sample conversation data of the data samples in the training sample set according to the sample information state of the data samples in the training sample set; The trained network model is used as the preset network model.

5. The method according to claim 4, characterized in that The training samples include data samples of multiple rounds of training; The step of establishing a network model to be trained includes: Establish the initial network parameters and initial reward function of the network model to be trained; The method of training the network parameters of the network model to be trained using the data samples in the training sample set based on the reinforcement learning algorithm includes: According to the reinforcement learning algorithm, the initial network parameters of the network model to be trained are trained using the data samples of any round of training to obtain sample conversation data; Updating the initial reward function according to the sample conversation data to obtain target network parameters and target reward function of the network model to be trained; Determine whether the training results are the expected results; If yes, the trained network model is used as the preset network model; If not, according to the reinforcement learning algorithm, the target network parameters in the network model to be trained are trained using the data samples of the next round of training to obtain sample conversation data, and the step of updating the initial reward function according to the sample conversation data is repeated until the trained network model is obtained.

6. A device for processing user input data combining RPA and AI, characterized in that: include: An acquisition module is used to obtain user input data to be processed; a data processing module, configured to process the user input data to obtain an attribute probability distribution and a matching probability distribution; wherein the attribute probability distribution is used to indicate the degree of similarity between the user input data and each attribute value under any attribute type, and the matching probability distribution is used to indicate the degree of matching between the user input data and each candidate result; A network model processing module, configured to input the information state of the user input data formed by the attribute probability distribution and the matching probability distribution into a preset network model and output conversation data; After obtaining the user input data to be processed, the method further includes: Determine whether this is the first round of processing of the user input data to be processed; If so, the method includes: According to the user input data, the results stored in the preset result candidate library are screened to obtain a number of candidate results; If not, the method further comprises: Process the candidate results, attribute probability distribution, and matching probability distribution of the previous round based on the user input data of this round to obtain the information state of the user input data composed of the attribute probability distribution of this round and the matching probability distribution, input it into the preset network model, and output conversation data; Each candidate result is stored in a preset candidate result library, which also stores the attribute type of each candidate result and the corresponding attribute value; the user input data is processed to obtain a matching probability distribution, including: Based on the attribute probability distribution of the user input data and each attribute value under any attribute type, as well as the attribute value of the attribute type of each candidate result, the degree of association between the user input data and each candidate result is determined, and the matching probability between the user input data and each candidate result is obtained; the matching probability of each candidate result constitutes the matching probability distribution.

7. An electronic device, characterized in that: include: at least one processor and memory; The memory stores computer-executable instructions; The at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs the method according to any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions, and when a processor executes the computer-executable instructions, the method according to any one of claims 1 to 5 is implemented.

Images