A method, system and electronic device for creating an RPA process based on natural language processing technology

Abstract

This invention proposes a method, system, and electronic device for creating RPA processes based on natural language processing technology, belonging to the technical field of human-computer interaction. The process of creating the process includes: receiving process information in a designer and performing sentence-by-sentence matching with data in a knowledge base; when the matching result is successful, generating corresponding XML process data; when the matching result is unsuccessful, performing sentence segmentation, part-of-speech tagging, and entity recognition to generate word vectors, and performing vector matching on the word vectors to obtain XML structure data; integrating the XML structure data to obtain XML process data; parsing the XML process data to obtain the corresponding RPA process and testing it; modifying the process parameters based on the test results to obtain the final process, and synchronously updating the data in the knowledge base. This invention solves the problem of new users finding it difficult to create processes and also reduces the time users spend creating processes.

Description

Technical Field

[0001] This invention relates to the field of human-computer interaction, and in particular to a method, system, and electronic device for creating RPA processes based on natural language processing technology. Background Technology

[0002] RPA technology stands for Robotic Process Automation. The corresponding application automates manual processes by mimicking the user's manual operation on a computer. Current RPA process creation involves first opening a process designer, designing the process based on familiarity with the business logic and the designer, dragging out corresponding nodes and connecting them according to the process sequence, modifying node parameters, testing the process, and fixing any bugs to complete the final RPA process creation.

[0003] It's easy to see that, based on the existing process creation steps, the selection of sequential nodes depends on the user's knowledge base. For beginners, it's difficult to design an independent process, and a high level of expertise is required for a complete and effective process. On the other hand, for experienced users, creating the same process involves repetitive operations, which often leads to tedium and reduced attention, resulting in a higher error rate. Summary of the Invention

[0004] Purpose of the invention: To propose a method, system, and electronic device for creating RPA processes based on natural language processing technology, so as to solve the above-mentioned problems existing in the prior art, automatically generate RPA process creation based on natural language, and overcome the cumbersome defects of traditional RPA process creation.

[0005] Technical Solution: Firstly, a method for creating RPA processes based on natural language processing technology is proposed. The process of creating an RPA process includes:

[0006] The designer accepts process information input from the user and performs sentence-by-sentence matching between the process information and the data in the knowledge base.

[0007] When the matching result of the whole sentence is successful, the corresponding XML process data is generated; when the matching result of the whole sentence is unsuccessful, the process information is processed by sentence segmentation, part-of-speech tagging and entity recognition to generate the corresponding word vectors, and vector matching is performed on the word vectors to obtain the corresponding XML structure data. At the same time, the XML structure data is integrated to obtain the XML process data.

[0008] Parse the XML process data to obtain the corresponding RPA process, and then test it;

[0009] Based on the test results, the process parameters are modified to obtain the final RPA process, and the data in the knowledge base is updated synchronously.

[0010] In some possible implementations of the first aspect, during the process of matching the process information with the data in the knowledge base, the similarity of the whole sentence is calculated to determine the match. When the calculated value is less than a preset threshold, it indicates that the match is successful, and after the match is successful, the complete formatted XML process data is returned.

[0011] The similarity matching expression is:

[0012]

[0013] In the formula, s and t represent the two strings to be compared; d(i,j) represents the minimum edit distance required to edit the first i characters of s into the first j characters of t; cost represents the cost required to replace s[i] with t[j].

[0014] When the matching result for the entire sentence fails, the process of obtaining XML process data includes:

[0015] The process information is segmented into words to obtain independent word units;

[0016] The word units are tagged with part-of-speech tags and recognized as entities. Using the results of part-of-speech tagging and entity recognition, word unit vectorization is performed through one-hot encoding to obtain the corresponding word vectors.

[0017] In some possible implementations of the first aspect, for word vectors in a dispersed state, the XML structure data is integrated using the bpm rule to obtain XML flow data.

[0018] In addition, when the final RPA process is obtained, the corresponding data information is transmitted to the server asynchronously, and after being parsed into vectors, it is stored in the knowledge base.

[0019] In some possible implementations of the first aspect, the process of creating an RPA process on the user side includes:

[0020] Enable the AI ​​workflow design function in the designer;

[0021] Enter the process information on the corresponding data input interface;

[0022] After completing the process information input, click the "Confirm" button to confirm the input information;

[0023] Obtain the corresponding process after processing information and execute tests;

[0024] Adjust parameters based on test results;

[0025] When the test results meet the user's requirements, save the corresponding RPA process and synchronize the corresponding process information to the knowledge base.

[0026] Secondly, a system for creating RPA processes based on natural language processing technology is proposed to implement the method of creating RPA processes based on natural language processing technology. The system includes: a knowledge base, a data reading module, a sentence comparison module, a recognition and judgment module, a first XML process data generation module, a second XML process data generation module, an XML process parsing module, a testing module, a fine-tuning module, and an update module.

[0027] The system comprises the following modules: a knowledge base for storing RPA process information; a data reading module for receiving process information from the user; a sentence-by-sentence matching module for matching process information with data in the knowledge base; a recognition and judgment module for judging the matching result of the sentence-by-sentence matching module; a first XML process data generation module for generating corresponding XML process data when the sentence-by-sentence matching result of the recognition and judgment module is successful; a second XML process data generation module for generating corresponding XML process data through sentence segmentation, part-of-speech tagging, entity recognition processing, word vector matching, and XML structure data integration operations when the sentence-by-sentence matching result of the recognition and judgment module is unsuccessful; an XML process parsing module for parsing the XML process data to obtain the corresponding RPA process; a testing module for testing the RPA process; a fine-tuning module for adjusting RPA process parameters based on the test results; and an update module for updating the knowledge base with the execution results of the fine-tuning module.

[0028] Thirdly, embodiments of the present invention provide an electronic device, including a processor, a memory, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, it implements the above-described method steps for creating an RPA process based on natural language processing technology.

[0029] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method steps provided in the above embodiments for performing an RPA process based on natural language processing technology.

[0030] Fifthly, embodiments of the present invention provide a computer program product, the product including a computer program, which, when executed by a processor, implements method steps for creating an RPA process based on natural language processing technology.

[0031] Beneficial Effects: This invention proposes a method, system, and electronic device for creating RPA processes based on natural language processing technology. By transforming the RPA process creation method, it reduces complaints about time costs. Simultaneously, it introduces a feedback mechanism, where the success rate increases with the number of times a user uses the process. This not only solves the problem of new users finding it difficult to create processes but also reduces the time users spend creating processes. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 Generate a data processing flowchart for the existing RPA process.

[0034] Figure 2 This is a data processing flowchart for generating RPA according to a certain embodiment of the present invention.

[0035] Figure 3 This is a data processing flowchart for a platform in one embodiment of the present invention.

[0036] Figure 4 This is a flowchart of the data processing on the user side of a certain embodiment of the present invention. Detailed Implementation

[0037] In the following description, numerous specific details are set forth in order to provide a more thorough understanding of the invention. However, it will be apparent to those skilled in the art that the invention can be practiced without one or more of these details. In other instances, certain technical features well-known in the art have not been described in order to avoid obscuring the invention.

[0038] With the development of intelligent technologies, Robotic Process Automation (RPA) technology has also made significant progress. To meet the need for automated task execution, current technologies generally rely on manual creation of different nodes during RPA process creation. This involves creating corresponding process nodes based on human experience and completing the final process creation through node connections and parameter modifications. It's easy to see that current technologies overly depend on human experience, making them unfriendly to newcomers. However, similar processes are often repetitive and tedious, leading experienced users to neglect them due to excessive repetition, thus increasing the error rate.

[0039] This application addresses the aforementioned problems by proposing a method, system, and electronic device for creating RPA processes based on natural language processing technology. By automating the process creation, it reduces reliance on human operators, automates the conversion of natural language into system processes, and thus improves the efficiency of RPA process creation. For example, when a user inputs "buy xx mobile phone from xx platform," the system can automatically parse the corresponding natural language instruction, thereby creating an RPA process. Simultaneously, it provides feedback on the execution results to assist in optimizing the current analysis method and improve the success rate of subsequent automatic RPA process creation.

[0040] Example 1

[0041] In one embodiment, a method for creating an RPA process based on natural language processing technology is proposed. This method includes the following steps: First, the process information is received in the designer and matched with data in the knowledge base using whole-sentence matching. Second, corresponding operations are performed based on the recognition results; that is, when the whole-sentence matching result is successful, corresponding XML process data is generated; when the whole-sentence matching result is unsuccessful, sentence segmentation, part-of-speech tagging, and entity recognition are performed to generate word vectors, and vector matching is performed on the word vectors to obtain XML structure data. After integrating the XML structure data, XML process data is obtained. Next, the XML process data is parsed to obtain the corresponding RPA process, and testing is performed. Finally, the process parameters are modified based on the test results to obtain the final process, and the data in the knowledge base is updated synchronously.

[0042] Specifically, during the sentence recognition process, sentence matching is completed through similarity calculation. In one embodiment, the edit distance algorithm is used to calculate sentence similarity matching, and the corresponding expression is:

[0043]

[0044] In the formula, s and t represent the two strings to be compared; d(i,j) represents the minimum edit distance required to edit the first i characters of s into the first j characters of t; cost represents the cost required to replace s[i] with t[j].

[0045] When the edit distance is less than the set threshold, it indicates a successful match, and the complete formatted XML data is returned for subsequent parsing. The threshold used for comparison can be adjusted based on the actual sentence length.

[0046] When the whole sentence matching result does not meet the preset threshold, the matching result of the whole sentence is recognized as failed, and then word segmentation, part-of-speech tagging, and entity recognition are performed to obtain the corresponding word vectors. In one embodiment, the word segmentation function uses the Hidden Markov Model (HMM). For the user input "Go to xx platform to buy xx mobile phone", it is segmented into "Go", "xx", "platform", "buy", "xx mobile phone"

[0047] Specifically, the process of performing word segmentation includes: First, construct a state set, an observation set, transition probabilities, and emission probabilities. Then, based on the given observation sequence, obtain the final state sequence through initialization of states, recursion, obtaining the maximum probability, and backtracking operations. Finally, perform word segmentation using the final state sequence. In one embodiment, construct a state set S, an observation set O, transition probabilities \(a_{ij}\), emission probabilities \(b_j(k)\), and initial state probabilities \(\pi_i\). In Chinese word segmentation, the state usually corresponds to the beginning B of a word, the middle M of a word, the end E of a word, and a single character forming a word S. The observation set O corresponds to all possible characters. The transition probability \(a_{ij}\) represents the probability of transitioning from state i to state j. For example, the probability of transitioning from the beginning state B of a word to the middle state M of a word is denoted as \(a_{BM}\). The emission probability \(b_j(k)\) represents the probability of observing symbol k in state j. For example, the probability of observing the character "中" in the beginning state B of a word is denoted as \(b_B(\text{"中"})\). The initial state probability \(\pi_i\) represents the probability of each state at the initial moment.

[0048] Among them, the expression corresponding to the transition probability \(a_{ij}\) is:

[0049] \(a_{ij} = P(i_{t + 1} = q_j|i_t = q_i),\quad1\leq i,j\leq N\)

[0050] In the formula, N represents the number of states; \(i_t\) represents the state at time t; \(i_{t + 1}\) represents the state at time t + 1.

[0051] The expression corresponding to the emission probability \(b_j(k)\) is:

[0052] \(b_j(k) = P(o_t = v_k|i_t = q_j),\quad1\leq j\leq N, k\in V\)

[0053] In the formula, V represents the observation set; \(o_t\) represents the observed value at time t.

[0054] The expression corresponding to the initial state probability \(\pi_i\) is:

[0055] \(pi_i = P(i_1 = q_i),\quad1\leq i\leq N\)

[0056] When performing word segmentation, based on the given observation sequence \(O\), obtain the most likely state sequence \(I\), and use the state sequence \(I\) to perform sentence-level word segmentation. First, initialize each state. Second, obtain \(\delta_t(i)\) through recursion. Third, when reaching the termination condition, obtain the maximum probability \(P^*\). Fourth, perform backtracking. Finally, obtain the state sequence and use the state sequence for word segmentation.

[0057] Among them, in the initialization process, for each state \((q_i)\), calculate \((\delta_1(i)=\pi_i b_i(o_1))\); in the recursion process, for \((t = 2, 3, \ldots, T)\), calculate \((\delta_t(i)=\max_{1\leq j\leq N}[\delta_{t - 1}(j)a_{ji}]b_i(o_t))\); the calculation expression for the maximum probability \(P^*\) is:

[0058] \(P^*=\max_{1\leq i\leq N}\delta_T(i)\); in the backtracking process, start from \(T\), for each \(t\), find \(i_t^*\) such that \((\delta_t(i_t^)=\max_j[\delta_{t - 1}(j)a_{ji}])\), until \((t = 1)\); finally, obtain the state sequence \(I^* = i_1^*, i_2^*, \ldots, i_T^*\).

[0059] In the above symbols, \(\max\) represents taking the maximum value; \(\delta\) represents the integral operation; \(\ldots\) represents omission; \(\quad\) represents generating a space; \(\leq\) represents less than or equal to.

[0060] Perform词性标注and entity recognition on the individual lexical units obtained after word segmentation. First, use a pre-trained model to perform the tasks of词性标注and named entity recognition on the input text; subsequently, obtain the词性标签and entity category information output by the model; finally, use one-hot encoding to convert the词性标签and entity category information into vector representations. In a preferred embodiment, during the process of vectorizing word segmentation, a pre-trained BERT model is used to perform词性标注and named entity recognition on the input text. For example, for "Go to the xx platform to buy an xx mobile phone", "go" is recognized as a verb, "xx platform" is recognized as a proper noun, "platform" is recognized as a noun, "buy" is recognized as a verb, and "xx mobile phone" is recognized as a proper noun.

[0061] Finally, after the process test is correct, save the process, and the system will asynchronously upload the process information to the server and parse it into a vector and save it in the knowledge base.

[0062] In one embodiment, such as Figure 2 As shown, when a user's requirement is "to buy xx mobile phone from xx website", the user first enters the AI ​​process design function of the designer and inputs the corresponding process information. Then, the received process information is matched with the data in the knowledge base. When the match is successful, the XML data is directly returned to the user for process parsing. When the match is unsuccessful, the process information is segmented into words, and the segmented words are identified as verbs, nouns, proper nouns, etc. through part-of-speech tagging and entity recognition, thereby generating word vectors. Then, the word vectors generated by the process information are matched in the knowledge base to generate corresponding XML data. Then, according to the bpm2.0 specification, the scattered XML data are merged into a process. Finally, the process is tested. When the test results show errors, the process is modified. After the test is successful, the process is saved, and the process information in the knowledge base is updated.

[0063] This embodiment's method for automatically creating RPA processes effectively overcomes the cumbersome defects of existing technologies during the creation process. It also addresses the problem of new users finding it difficult to create processes independently. Through natural language analysis and instruction conversion, it not only solves the problem of new users struggling to create processes but also reduces the time required for process creation. Furthermore, this embodiment introduces a feedback mechanism, which increases the success rate of automatic RPA process creation in tandem with the number of times the user uses the system.

[0064] Example 2

[0065] In one embodiment, during the RPA process creation process, the interactive platform and the user client execute the process creation process, and the interactive platform analyzes and recognizes natural language to generate the RPA process, as follows: Figure 3 As shown, it includes:

[0066] Step 1: Receive process information input from the user;

[0067] Step 2: Compare the process information with the stored information in the knowledge base. If the comparison is successful, generate the corresponding XML structure data and proceed to step 8. If the comparison is unsuccessful, proceed to step 3.

[0068] Step 3: Perform word segmentation on the process information to obtain individual word units;

[0069] Step 4: Perform part-of-speech tagging and entity recognition on the vocabulary units to generate corresponding word vectors;

[0070] Step 5: Perform vector matching between word vectors and stored information in the knowledge base; In a preferred embodiment, during the vector matching process, the similarity between vectors is obtained according to the K-nearest neighbor algorithm, then the node or component with the highest similarity is found, and the corresponding data is integrated into an XML structure data format and returned to the Studio client;

[0071] Step 6: Generate XML structure data based on the vector matching results;

[0072] Step 7: Merge the XML structure data according to the bpm2.0 specification;

[0073] Step 8: Parse the XML structure data and generate the corresponding data processing flow;

[0074] Step 9: Test the generated data processing flow and adjust the flow parameters based on the test results;

[0075] Step 10: The final tested process is automatically created as the result, and the relevant process information is synchronized to the knowledge base. In a preferred embodiment, after the process test is successful, the process is saved, and the interactive platform system asynchronously uploads the process information to the server, parses it into a vector, and saves it into the knowledge base.

[0076] This embodiment transforms the RPA process creation process into an automatic conversion based on user-input requirement text. From the system platform level, compared to the drawbacks of existing technologies that require users to manually create process nodes, this embodiment greatly reduces reliance on human operation and improves the efficiency of automated process creation. Furthermore, reducing user decision-making at process nodes makes the final generated process more objective and accurate.

[0077] Example 3

[0078] In one embodiment, such as Figure 4 As shown, the RPA process executed on the user side includes the following steps:

[0079] Step 1: Open the designer;

[0080] Step 2: Enter the process information in the corresponding text box in the designer, and submit after confirmation;

[0081] In one embodiment, the user opens the AI ​​process design function of the designer and enters process information in the dialog box, such as "go to xx website to buy xx mobile phone". After confirming that the information is correct, the user clicks the confirmation button.

[0082] Step 3: Obtain the automatically generated RPA process information;

[0083] Step 4: Test the obtained RPA process information;

[0084] Step 5: Adjust the parameters of the RPA process information based on the test results; if the test results do not meet the requirements, perform the test again; otherwise, jump directly to step 6.

[0085] Step 6: When the test results meet the requirements, enter the corresponding save command to complete the saving of the RPA process information.

[0086] This embodiment proposes a method for automatically generating RPA processes from the user's perspective. It achieves automatic creation of RPA processes based on natural language. Compared with existing technologies, it effectively reduces the impact of user selection on process nodes. By reducing reliance on users, it effectively reduces the strong requirement for users' professional knowledge and overcomes the shortcomings of novices creating the required processes.

[0087] Example 4

[0088] In one embodiment, a system for creating RPA processes based on natural language processing technology is proposed to implement a method for creating RPA processes based on natural language processing technology. The system includes: a knowledge base, a data reading module, a sentence comparison module, an identification and judgment module, a first XML process data generation module, a second XML process data generation module, an XML process parsing module, a testing module, a fine-tuning module, and an update module.

[0089] The system comprises the following modules: a knowledge base for storing RPA process information; a data reading module for receiving process information from the user; a sentence-by-sentence matching module for matching process information with data in the knowledge base; a recognition and judgment module for judging the matching result of the sentence-by-sentence matching module; a first XML process data generation module for generating corresponding XML process data when the sentence-by-sentence matching result of the recognition and judgment module is successful; a second XML process data generation module for generating corresponding XML process data through sentence segmentation, part-of-speech tagging, entity recognition processing, word vector matching, and XML structure data integration operations when the sentence-by-sentence matching result of the recognition and judgment module is unsuccessful; an XML process parsing module for parsing the XML process data to obtain the corresponding RPA process; a testing module for testing the RPA process; a fine-tuning module for adjusting RPA process parameters based on the test results; and an update module for updating the knowledge base with the execution results of the fine-tuning module.

[0090] In one embodiment, during the creation of the RPA process, the process first uses a data reading module to receive process information input from the user; second, a sentence comparison module performs sentence matching between the process information and data in the knowledge base, and a recognition and judgment module compares the results. When the sentence matching result of the recognition and judgment module is successful, the first XML process data generation module generates the corresponding XML process data; when the sentence matching result of the recognition and judgment module is unsuccessful, the second XML process data generation module performs sentence segmentation, part-of-speech tagging, entity recognition processing, word vector matching, and XML structure data integration operations to generate the corresponding XML process data; third, the XML process parsing module parses the XML process data to obtain the corresponding RPA process, and the testing module performs RPA process testing; fourth, based on the test results of the testing module, the fine-tuning module adjusts the RPA process parameters; finally, the final RPA process information is obtained, and the update module stores the final RPA process information in the knowledge base.

[0091] Example 5

[0092] In one embodiment, an apparatus for creating an RPA process based on natural language processing technology is provided. The apparatus includes a processor and a memory storing computer program instructions. The processor reads and executes the computer program instructions to implement a method for creating an RPA process based on natural language processing technology.

[0093] Specifically, electronic devices include computers in various forms, including but not limited to: laptops, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframes, and other suitable computers. Additionally, electronic devices can also be presented as various forms of mobile devices, such as cellular phones, smartphones, wearable devices, and other computer devices.

[0094] Optionally, the device for automatically generating electronic reports also includes components for implementing different functions, and these components are all connected to the I / O interface via a bus. These functional components include, but are not limited to: input units, output units, storage units, and communication units. Input units include, for example, a mouse and keyboard; output units include, for example, different types of displays and speakers; storage units include, for example, disks and optical discs; and communication units include, for example, network cards, modems, and wireless transceivers.

[0095] The communication unit is used to realize data transmission between electronic devices and other target objects. Optionally, it can exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0096] Example 6

[0097] In one embodiment, a computer program is also provided. When executed by a processor, the computer program implements the various processes of the above-described method embodiment for creating an RPA process based on natural language processing technology, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0098] Example 7

[0099] In one embodiment, a computer-readable storage medium is also provided, on which a computer program is stored. When executed by a processor, the computer program implements the various processes of the above-described method embodiment for creating an RPA process based on natural language processing technology, and achieves the same technical effect. To avoid repetition, it will not be described again here. The computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0100] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0101] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0102] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0103] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0104] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0105] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0106] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0107] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0108] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0109] As described above, although the invention has been shown and described with reference to specific preferred embodiments, it should not be construed as limiting the invention itself. Various changes in form and detail may be made without departing from the spirit and scope of the invention as defined in the appended claims.

Claims

1. A method for creating RPA processes based on natural language processing technology, characterized in that, The process of creating an RPA process includes: The designer accepts process information input from the user and performs sentence-by-sentence matching between the process information and the data in the knowledge base. When the matching result of the whole sentence is successful, the corresponding XML process data is generated; when the matching result of the whole sentence is unsuccessful, the process information is processed by sentence segmentation, part-of-speech tagging and entity recognition to generate the corresponding word vectors, and vector matching is performed on the word vectors to obtain the corresponding XML structure data. At the same time, the XML structure data is integrated to obtain the XML process data. Parse the XML process data to obtain the corresponding RPA process, and then test it; Based on the test results, modify the process parameters to obtain the final RPA process, and update the data in the knowledge base synchronously. During the process of matching the process information with the data in the knowledge base, the similarity of the whole sentence is calculated to determine the match. When the calculated value is less than the preset threshold, it indicates that the match is successful, and complete formatted XML process data is returned after the match is successful. The similarity matching expression is: In the formula, s and t represent the two strings to be compared; This represents the minimum edit distance required to edit the first i characters of s into the first j characters of t; Indicates will Replace with The cost required.

2. The method for creating an RPA process based on natural language processing technology according to claim 1, characterized in that, When the matching result for the entire sentence fails, the process of obtaining XML process data includes: The process information is segmented into words to obtain independent word units; The word units are tagged with part-of-speech tags and recognized as entities. Using the results of part-of-speech tagging and entity recognition, word unit vectorization is performed through one-hot encoding to obtain the corresponding word vectors.

3. The method for creating an RPA process based on natural language processing technology according to claim 1, characterized in that, For word vectors in a scattered state, the BPM rule is used to integrate the XML structure data to obtain XML process data.

4. The method for creating an RPA process based on natural language processing technology according to claim 1, characterized in that, When the final RPA process is obtained, the corresponding data information is transmitted to the server asynchronously, and after being parsed into vectors, it is stored in the knowledge base.

5. The method for creating an RPA process based on natural language processing technology according to claim 1, characterized in that, The process of creating an RPA process on the user's end includes: Enable the AI ​​workflow design function in the designer; Enter the process information on the corresponding data input interface; After completing the process information input, click the "Confirm" button to confirm the input information; Obtain the corresponding process after processing information and execute tests; Adjust parameters based on test results; When the test results meet the user's requirements, save the corresponding RPA process and synchronize the corresponding process information to the knowledge base.

6. A system for creating RPA processes based on natural language processing technology, used to implement the method for creating RPA processes based on natural language processing technology as described in any one of claims 1-5, characterized in that, Includes the following modules: The knowledge base is configured to store RPA process information; The data reading module is configured to accept process information input from the user terminal; The whole-sentence comparison module is configured to perform whole-sentence matching between the process information and the data in the knowledge base; The identification and judgment module is set to judge the matching results of the whole sentence comparison module; The first XML process data generation module is set to generate corresponding XML process data when the whole sentence matching result of the recognition and judgment module is successful. The second XML process data generation module is configured to generate corresponding XML process data when the whole sentence matching result of the recognition judgment module is a recognition failure, through sentence segmentation, part-of-speech tagging, entity recognition processing, word vector matching, and XML structure data integration operations. The XML process parsing module is configured to parse XML process data to obtain the corresponding RPA process; The testing module is configured to test the RPA process; The fine-tuning module is configured to adjust RPA process parameters based on the test results of the RPA process. The update module is configured to update the knowledge base with the execution results of the fine-tuning module.

7. An electronic device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, it implements the method steps for creating an RPA process based on natural language processing technology as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, and when executed by a processor, the computer program implements the method steps for creating an RPA process based on natural language processing technology as described in any one of claims 1 to 5.

9. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method steps for creating an RPA process based on natural language processing technology as described in any one of claims 1 to 5.

Images