Text feature generation method, device, equipment and storage medium combining RPA and AI

Abstract

The present disclosure proposes a text feature generation method, device, equipment and storage medium combining RPA and AI, which is applied to an intelligent dialogue system supporting natural language processing NLP. The intelligent dialogue system can be registered and used by users. The method includes: obtaining dialogue text based on robot process automation RPA, determining the first word vector feature of the dialogue text based on artificial intelligence AI, determining the user's demand word vector intention, and processing the first word vector feature according to the demand word vector intention to obtain the second word vector feature of the dialogue text. Through the present disclosure, RPA combined with artificial intelligence AI can be used to pre-identify the user's demand word vector intention, so as to adjust and process the first word vector feature of the dialogue text, ensure the accuracy of word vector recognition and matching, effectively improve the accuracy of text feature generation, and improve the recall performance of the intelligent dialogue system.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to a text feature generation method, device, equipment and storage medium combining RPA (Robotic Process Automation) and AI (Artificial Intelligence). Background Art

[0002] Robotic Process Automation (RPA), referred to as Robotic Process Automation, uses specific "robot software" to simulate human operations on computers and automatically execute process tasks according to rules.

[0003] Artificial Intelligence (AI) is a technical science that studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence.

[0004] An intelligent dialogue system refers to a dialogue system that uses processing algorithms involved in artificial intelligence (AI) to process user question text to obtain response text corresponding to the question text. The user refers to the object that interacts with the intelligent dialogue system.

[0005] In the related art, when an intelligent dialogue system is used to recall approximate answers and intentions to user questions, the conversation text between the user and the intelligent dialogue system is usually used to train and fine-tune the artificial intelligence AI model.

[0006] In this way, the amount of data from a single user's conversation text is insufficient to fine-tune the artificial intelligence (AI) model, which affects the accuracy of the conversation text feature analysis and affects the recall performance of the intelligent conversation system. Summary of the invention

[0007] The present disclosure aims to solve one of the technical problems in the related art at least to some extent.

[0008] To this end, the purpose of the present disclosure is to propose a text feature generation method, device, electronic device and storage medium combining RPA and AI, which can use RPA combined with artificial intelligence AI to pre-identify the user's demand word vector intention, so as to adjust the first word vector feature of the dialogue text, ensure the accuracy of word vector recognition and matching, effectively improve the accuracy of text feature generation, and improve the recall performance of the intelligent dialogue system.

[0009] The text feature generation method combining RPA and AI proposed in the first aspect of the present disclosure includes: obtaining a conversation text based on robot process automation RPA; determining a first word vector feature of the conversation text based on artificial intelligence AI; determining the user's demand word vector intention; and processing the first word vector feature according to the demand word vector intention to obtain a second word vector feature of the conversation text.

[0010] In one embodiment, the second word vector feature is processed to obtain a sentence vector feature of the dialogue text, wherein the intelligent dialogue system processes the sentence vector feature based on the AI ​​to obtain a response text of the dialogue text.

[0011] In one embodiment, the determining the first word vector feature of the conversation text based on artificial intelligence AI includes: determining the first word vector of the conversation text based on the AI; determining the first word frequency corresponding to the first word vector based on a smoothed inverse word frequency algorithm; determining the first word weight of the first word vector based on the first word frequency, wherein the first word vector and the first word weight are collectively used as the first word vector feature.

[0012] In one embodiment, after determining the user's demand word vector intention, it also includes: processing the demand word vector intention based on the AI ​​to obtain multiple intention word vectors; performing principal component decomposition on the multiple intention word vectors to obtain common semantic information between the multiple intention word vectors; and determining a target intention word vector from the multiple intention word vectors based on the common semantic information.

[0013] In one embodiment, the first word vector feature is processed according to the required word vector intention to obtain the second word vector feature of the conversation text, including: if the target intention word vector and the first word vector match, the first word weight is increased to be the target word weight; and the first word vector and the target word weight are used as the second word vector feature.

[0014] In one embodiment, the processing of the second word vector feature to obtain the sentence vector feature of the conversation text includes: if the target intention word vector matches the first word vector, then describing the first word vector based on the target word weight to obtain the sentence vector feature of the conversation text.

[0015] In one embodiment, the first word vector feature is processed according to the required word vector intention to obtain the second word vector feature of the conversation text, including: if the target intention word vector does not match the first word vector, using the target intention word vector to update the first word vector; configuring the intention word weight for the target intention word vector; and using the target intention word vector and the intention word weight as the second word vector feature.

[0016] In one embodiment, the processing of the second word vector feature to obtain the sentence vector feature of the conversation text includes: if the target intention word vector does not match the first word vector, describing the target intention word vector based on the intention word weight to obtain the sentence vector feature of the conversation text.

[0017] In one embodiment, before configuring the intention word weight for the target intention word vector, it also includes: inputting the requirement word vector intention into the weight parsing model in the AI ​​to obtain the intention word weight corresponding to the target intention word vector.

[0018] The text feature generation device combining RPA and AI proposed in the second aspect of the embodiment of the present disclosure includes: an acquisition module, used to acquire a conversation text based on robot process automation RPA; a first determination module, used to determine a first word vector feature of the conversation text based on artificial intelligence AI; a second determination module, used to determine the user's demand word vector intention; and a first processing module, used to process the first word vector feature according to the demand word vector intention to obtain a second word vector feature of the conversation text.

[0019] In one embodiment, it also includes: a second processing module, used to process the second word vector feature to obtain the sentence vector feature of the dialogue text, wherein the intelligent dialogue system processes the sentence vector feature based on the AI ​​to obtain the response text of the dialogue text.

[0020] In one embodiment, the first determination module includes: a first determination submodule, used to determine the first word vector of the conversation text based on the AI; a second determination submodule, used to determine the first word frequency corresponding to the first word vector based on a smoothed inverse word frequency algorithm; a third determination submodule, used to determine the first word weight of the first word vector based on the first word frequency, wherein the first word vector and the first word weight are jointly used as the first word vector feature.

[0021] In one embodiment, it also includes: a third processing module, which is used to process the demand word vector intention based on the AI ​​to obtain multiple intention word vectors after determining the user's demand word vector intention; a fourth processing module, which is used to perform principal component decomposition on the multiple intention word vectors to obtain common semantic information between the multiple intention word vectors; and a third determination module, which is used to determine the target intention word vector from the multiple intention word vectors based on the common semantic information.

[0022] In one embodiment, the first processing module is specifically used to: if the target intention word vector and the first word vector match, increase the first word weight to be the target word weight; and use the first word vector and the target word weight as the second word vector features.

[0023] In one embodiment, the second processing module is specifically used to: if the target intention word vector matches the first word vector, then describe the first word vector based on the target word weight to obtain the sentence vector feature of the dialogue text.

[0024] In one embodiment, the first processing module is also used to: if the target intention word vector does not match the first word vector, use the target intention word vector to update the first word vector; configure the intention word weight for the target intention word vector; and use the target intention word vector and the intention word weight as the second word vector features.

[0025] In one embodiment, the second processing module is further used to: if the target intention word vector does not match the first word vector, describe the target intention word vector based on the intention word weight to obtain the sentence vector feature of the dialogue text.

[0026] In one embodiment, the first processing module is also used to: before configuring the intention word weight for the target intention word vector, input the requirement word vector intention into the weight parsing model in the AI ​​to obtain the intention word weight corresponding to the target intention word vector.

[0027] The third aspect embodiment of the present disclosure proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the text feature generation method combining RPA and AI proposed in the first aspect embodiment of the present disclosure is implemented.

[0028] The fourth aspect embodiment of the present disclosure proposes a non-temporary computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, it implements the text feature generation method combining RPA and AI as proposed in the first aspect embodiment of the present disclosure.

[0029] The advantages or beneficial effects of the above technical solution include at least: obtaining the conversation text based on robot process automation RPA, determining the first word vector feature of the conversation text based on artificial intelligence AI, determining the user's demand word vector intention, and processing the first word vector feature according to the demand word vector intention to obtain the second word vector feature of the conversation text. RPA combined with artificial intelligence AI can be used to pre-identify the user's demand word vector intention, so as to adjust the first word vector feature of the conversation text, ensure the accuracy of word vector recognition and matching, effectively improve the accuracy of text feature generation, and improve the recall performance of the intelligent dialogue system.

[0030] The above summary is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments and features described above, further aspects, embodiments and features of the present disclosure will be readily apparent by reference to the accompanying drawings and the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In the accompanying drawings, unless otherwise specified, the same reference numerals throughout the multiple drawings represent the same or similar parts or elements. These drawings are not necessarily drawn to scale. It should be understood that these drawings only depict some embodiments according to the present disclosure and should not be regarded as limiting the scope of the present disclosure.

[0032] Figure 1 It is a flowchart of a method for generating text features combining RPA and AI proposed in one embodiment of the present disclosure;

[0033] Figure 2 is a flowchart of a method for generating text features combining RPA and AI proposed in another embodiment of the present disclosure;

[0034] Figure 3 is a flowchart of a method for generating text features combining RPA and AI proposed in another embodiment of the present disclosure;

[0035] Figure 4 It is a structural diagram of a text feature generation device combining RPA and AI proposed in one embodiment of the present disclosure;

[0036] Figure 5 is a structural schematic diagram of a text feature generation device combining RPA and AI proposed in another embodiment of the present disclosure;

[0037] Figure 6A structural block diagram of an electronic device according to an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0038] Embodiments of the present disclosure are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present disclosure, and cannot be understood as limiting the present disclosure.

[0039] In the description of the present disclosure, the term "plurality" refers to two or more.

[0040] In the description of this disclosure, the term "robotic process automation (RPA)" refers to the automatic execution of process tasks on a computer according to rules through robotic application software.

[0041] In the description of this disclosure, the term "artificial intelligence AI" refers to the discipline that studies how computers can simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, planning, etc.), including both hardware-level and software-level technologies. Artificial intelligence hardware technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, and big data processing; artificial intelligence software technologies mainly include computer vision technology, speech recognition technology, natural language processing technology, as well as machine learning, deep learning, big data processing technology, knowledge graph technology, and other major directions.

[0042] In the description of the present disclosure, the term "dialogue text" refers to the content text of the interaction between the user and the robot. The content text of the interaction can be, for example, the user's inquiry text. For example, the dialogue text can specifically be the question text of the user inquiring about the product, or it can also be any other form of interactive content text, without limitation.

[0043] In the description of the present disclosure, the term "first word vector feature" refers to a semantic vector feature mapped to a vector space, which can be used to describe the semantics of a conversation text, and the first word vector feature can be used to capture the semantic similarity of words in the text.

[0044] In the description of the present disclosure, the term "second word vector feature" refers to the word vector feature that is matched and adjusted with the first word vector feature according to the user's demand word vector intention. The second word vector feature can be used for weighted processing to generate a sentence vector feature corresponding to the conversation text.

[0045] In the description of the present disclosure, the term “first word vector” refers to a representation vector generated after mapping the vocabulary in the conversation text in the vector space.

[0046] In the description of the present disclosure, the term “word frequency” refers to the frequency at which a word vector appears in a text, and the first word frequency refers to the frequency at which a first word vector appears in a text.

[0047] In the description of the present disclosure, the term "word weight" refers to the weight information of a word vector, and the first word weight is the weight information of the first word vector.

[0048] In the description of the present disclosure, the term "sentence vector feature" refers to a vector feature that can represent the semantics of a sentence and is obtained by weighting the second word vector feature.

[0049] In the description of the present disclosure, the term "response text" refers to text generated based on sentence vector features to respond to the text content in the dialogue text.

[0050] In the description of the present disclosure, the term "intent word vector" refers to a word vector obtained after mapping the user's demand word vector intent to a vector space. The demand word vector intent can be vectorized to obtain the intent word vector.

[0051] In the description of the present disclosure, the term "principal component decomposition" refers to converting multiple intent word vectors to obtain common semantic information between the multiple intent word vectors, so that the processed intent word vectors are unrelated intent word vectors.

[0052] In the description of the present disclosure, the term "target intent word vector" refers to an irrelevant intent word vector obtained after removing some intent word vectors containing common semantic information from multiple intent word vectors. There is no correlation between the target intent word vectors.

[0053] These and other aspects of the embodiments of the present disclosure will be apparent with reference to the following description and drawings. In these descriptions and drawings, some specific implementations of the embodiments of the present disclosure are specifically disclosed to represent some ways of implementing the principles of the embodiments of the present disclosure, but it should be understood that the scope of the embodiments of the present disclosure is not limited thereto. On the contrary, the embodiments of the present disclosure include all changes, modifications and equivalents that fall within the spirit and connotation of the appended claims.

[0054] The following describes a text feature generation method combining RPA and AI according to an embodiment of the present disclosure in conjunction with the accompanying drawings.

[0055] Figure 1 It is a flowchart of a method for generating text features combining RPA and AI proposed in one embodiment of the present disclosure.

[0056] In this embodiment, the text feature generation method combining RPA and AI is configured as a text feature generation device combining RPA and AI. In this embodiment, the text feature generation method combining RPA and AI can be configured in a text feature generation device combining RPA and AI. The text feature generation device combining RPA and AI can be set in a server, or can also be set in an electronic device, and the embodiments of the present disclosure are not limited to this.

[0057] This embodiment takes the text feature generation method combining RPA and AI as an example, which is configured in an electronic device, such as a smart phone, a tablet computer, a personal digital assistant, an e-book, and other hardware devices with various operating systems.

[0058] It should be noted that the execution subject of the embodiment of the present disclosure may be, for example, a central processing unit (CPU) in a server or electronic device in terms of hardware, or may be, for example, a related background service in a server or electronic device in terms of software, without limitation.

[0059] In addition, the "generation" in the embodiments of the present disclosure refers to the text feature generation process that combines robotic process automation RPA and artificial intelligence AI. That is to say, the text feature generation process is a fully automated generation process, and the text feature generation process can also be combined with artificial intelligence AI to realize automated response and intent recall in natural language processing (Natural Language Processing, NLP).

[0060] The present disclosure can be specifically applied to an intelligent dialogue system supporting natural language processing (NLP), which can be registered and used by users. Natural language processing (NLP) is a field of computer science, artificial intelligence, and linguistics that focuses on the interaction between computers and human (natural) language.

[0061] For example, in the embodiment of the present disclosure, the text feature generation process based on the full-process automation can realize the full-process automation execution of obtaining conversation text based on robotic process automation RPA, determining the first word vector feature of the conversation text based on artificial intelligence AI, determining the user's demand word vector intention, and processing the first word vector feature according to the demand word vector intention to obtain the second word vector feature of the conversation text.

[0062] like Figure 1 As shown, the text feature generation method combining RPA and AI includes:

[0063] The text feature generation method combining RPA and AI in the embodiments of the present disclosure can be applied to an intelligent dialogue system that supports natural language processing (NLP). The intelligent dialogue system is available for user registration and use, wherein the user refers to an object that interacts with the intelligent dialogue system, and the intelligent dialogue system refers to a dialogue system that uses the processing algorithms involved in artificial intelligence (AI) to process the user's question text to obtain an answer text corresponding to the question text.

[0064] S101: Obtaining conversation text based on robotic process automation (RPA).

[0065] Among them, Robotic Process Automation (RPA) refers to the automatic execution of process tasks on computers according to rules through robot application software.

[0066] Among them, the dialogue text refers to the content text of the interaction between the user and the robot. The interactive content text can be, for example, the user's inquiry text. For example, the dialogue text can specifically be the question text of the user asking about the product, or it can also be any other form of interactive content text, without limitation.

[0067] That is to say, an application scenario in the embodiments of the present disclosure may be, for example, obtaining the user's inquiry question text about the product based on robotic process automation (RPA). For example, the inquiry question may be "What is the affordable wine?" The obtained inquiry question text is used as the dialogue text, and then the obtained dialogue text may be subjected to text feature generation combining RPA and AI. The text feature generation method combining RPA and AI described in the embodiments of the present disclosure may also be applied to any other possible text feature generation scenario, without limitation.

[0068] In the embodiment of the present disclosure, when obtaining conversation text based on robot process automation (RPA), an application data interface can be pre-configured, and the RPA robot receives the content text of the interaction between the user and the robot through the application data interface according to the pre-set software operation process, and uses the obtained content text of the interaction between the user and the robot as the conversation text.

[0069] In other embodiments, the conversation audio of the user interacting with the robot can be obtained based on the robotic process automation (RPA), and then the speech recognition model can be used to perform text recognition and extraction on the interaction audio to extract the interaction content text corresponding to the conversation audio, and the obtained interaction content text can be used as the conversation text. Alternatively, a data transmission interface can be provided on a text feature generation device combining RPA and AI, and conversation text transmitted by other electronic devices can be received via the data transmission interface as the obtained conversation text. Alternatively, any other possible method based on robotic process automation (RPA) can be used to obtain the conversation text, and there is no limitation on this.

[0070] S102: Determine the first word vector feature of the conversation text based on artificial intelligence AI.

[0071] The first word vector feature refers to a semantic vector feature mapped to a vector space. The semantic feature can be used to describe the semantics of a conversation text. The first word vector feature can be used to capture the semantic similarity of words in a text.

[0072] After acquiring the conversation text based on the robotic process automation (RPA) as described above, the disclosed embodiment may determine the first word vector feature of the conversation text based on artificial intelligence (AI).

[0073] In the disclosed embodiment, when determining the first word vector feature of a conversation text based on artificial intelligence AI, a smooth inverse frequency algorithm (Smooth Inverse Frequency, SIF) can be used to perform word frequency analysis and weight analysis on the conversation text to obtain the word vector feature output by the smooth inverse frequency algorithm for describing the semantics of the conversation text, and the word vector feature output by the smooth inverse frequency algorithm is used as the first word vector feature of the conversation text.

[0074] In other embodiments, a pre-trained language analysis model including an input layer, a mapping layer, and an output layer may be used to analyze and process the dialogue text, a Huffman tree of the vocabulary in the dialogue text may be established in the output layer, and the final output result of the output layer may be obtained as the first word vector feature of the dialogue text, or any other method based on artificial intelligence AI may be used to determine the first word vector feature of the dialogue text, without limitation.

[0075] S103: Determine the user's demand word vector intention.

[0076] Among them, the user refers to the object that interacts with the intelligent dialogue robot through the dialogue text, and the user's demand word vector intention refers to the word feature intention set in advance according to the user's customized semantic association between words. The demand word vector intention can be used to customize the word vector of the dialogue text for closer processing.

[0077] In the disclosed embodiment, when determining the user's demand word vector intention, the semantic associations between words in the user-defined dialogue text can be obtained in advance, and then the semantic associations between the words can be encoded using a semantic coding model to zoom in on the word vectors containing the semantic associations to determine the user's demand word vector intention.

[0078] S104: Process the first word vector feature according to the required word vector intention to obtain a second word vector feature of the conversation text.

[0079] Among them, the second word vector feature refers to the word vector feature that is matched and adjusted with the first word vector feature according to the user's demand word vector intention. The second word vector feature can be used for weighted processing to generate a sentence vector feature corresponding to the dialogue text.

[0080] After determining the first word vector feature of the conversation text based on artificial intelligence AI and determining the user's demand word vector intention, the embodiment of the present disclosure can process the first word vector feature according to the demand word vector intention to obtain the second word vector feature of the conversation text.

[0081] In the disclosed embodiment, when processing the first word vector feature according to the required word vector intention to obtain the second word vector feature of the conversation text, the smoothed inverse word frequency algorithm SIF can be tuned according to the required word vector intention, and then the first word vector feature can be matched and adjusted using the tuned smoothed inverse word frequency algorithm SIF to obtain the vector feature output by the tuned smoothed inverse word frequency algorithm as the second word vector feature of the conversation text.

[0082] In other embodiments, a deep learning model can be pre-trained, which can perform the task of matching and adjusting word vector features, and the deep learning model can be targeted and adaptively tuned according to the required word vector intention to obtain a tuned deep learning model, and then the tuned deep learning model is used to match and adjust the first word vector feature to obtain the output result of the deep learning model as the second word vector feature of the conversation text, or the first word vector feature can be processed in any other way according to the required word vector intention to obtain the second word vector feature of the conversation text, without limitation.

[0083] In this embodiment, by obtaining the conversation text based on robot process automation RPA, determining the first word vector feature of the conversation text based on artificial intelligence AI, determining the user's demand word vector intention, and processing the first word vector feature according to the demand word vector intention to obtain the second word vector feature of the conversation text, RPA can be combined with artificial intelligence AI to pre-identify the user's demand word vector intention, so as to adjust the first word vector feature of the conversation text, ensure the accuracy of word vector recognition and matching, effectively improve the accuracy of text feature generation, and improve the recall performance of the intelligent dialogue system.

[0084] Figure 2 It is a flowchart of a method for generating text features combining RPA and AI proposed in another embodiment of the present disclosure.

[0085] like Figure 2 As shown, the text feature generation method combining RPA and AI includes:

[0086] S201: Acquire the conversation text based on the Robotic Process Automation (RPA).

[0087] The description of S201 can be referred to the above embodiment, and will not be repeated here.

[0088] S202: Determine the first word vector of the conversation text based on AI.

[0089] The first word vector refers to a representation vector generated after mapping the words in the dialogue text in the vector space.

[0090] In the disclosed embodiment, after acquiring the conversation text based on the robotic process automation RPA, the first word vector of the conversation text can be determined based on the artificial intelligence AI, and the conversation text can be segmented and vectorized using a pre-trained language representation model to obtain the word vector output by the pre-trained language representation model, and the word vector output by the pre-trained language representation model can be used as the first word vector of the conversation text.

[0091] S203: Determine a first word frequency corresponding to the first word vector based on a smoothed inverse word frequency algorithm.

[0092] Among them, word frequency refers to the frequency of word vectors appearing in the text, and first word frequency refers to the frequency of the first word vector appearing in the text.

[0093] In an embodiment of the present disclosure, after the conversation text is segmented and vectorized to obtain a first word vector for determining the conversation text, the frequency of occurrence of each first word vector can be statistically processed based on a smoothed inverse word frequency algorithm to obtain a first word frequency corresponding to the first word vector.

[0094] S204: Determine a first word weight of the first word vector according to the first word frequency, wherein the first word vector and the first word weight are jointly used as a first word vector feature.

[0095] Among them, word weight refers to the weight information of the word vector, and the first word weight is the weight information of the first word vector.

[0096] In the embodiment of the present disclosure, after determining the first word vector of the conversation text based on artificial intelligence AI and determining the first word frequency corresponding to the first word vector based on the smoothed inverse word frequency algorithm, the first word weight of the first word vector can be determined according to the first word frequency.

[0097] In the embodiment of the present disclosure, when determining the first word weight of the first word vector according to the first word frequency, the weight calculation formula can be used: Calculate the first word weight of the first word vector, where a is a constant. The parameter range of the constant a can be adjusted to make the calculation result close to the optimal result. p(w) represents the first word frequency of the first word vector, so as to obtain the weight calculation result of the weight calculation formula as the first word weight corresponding to the first word vector.

[0098] In the disclosed embodiment, after determining the first word weight of the first word vector according to the first word frequency using the weight calculation formula, the first word vector and the first word weight can be used together as the first word vector feature.

[0099] In this embodiment, the first word vector of the dialogue text is determined based on artificial intelligence AI, the first word frequency corresponding to the first word vector is determined based on the smoothed inverse word frequency algorithm, and the first word weight of the first word vector is determined according to the first word frequency, so that the first word weight value of the first word vector can be determined. Since the first word weight can be used to perform weighted processing on the first word vector to obtain a sentence vector feature, the accuracy of the generated sentence vector feature can be improved, thereby assisting in generating a more accurate response text corresponding to the dialogue text.

[0100] S205: Determine the user's demand word vector intention.

[0101] S206: Process the first word vector feature according to the required word vector intention to obtain a second word vector feature of the conversation text.

[0102] The description of S205 - S206 can be referred to the above embodiment for examples, and will not be repeated here.

[0103] S207: Process the second word vector feature to obtain a sentence vector feature of the dialogue text, wherein the intelligent dialogue system processes the sentence vector feature based on AI to obtain a response text of the dialogue text.

[0104] Among them, the sentence vector feature refers to the vector feature that can represent the semantics of the sentence obtained by weighted processing of the second word vector feature.

[0105] The response text refers to the text generated based on the sentence vector features to respond to the text content in the dialogue text.

[0106] Among them, the intelligent dialogue system refers to a dialogue system that uses the processing algorithms involved in artificial intelligence AI to process the user's question text to obtain the response text corresponding to the question text. The intelligent dialogue system can process sentence vector features based on AI to obtain the response text of the dialogue text.

[0107] After determining the user's demand word vector intention as described above and processing the first word vector feature according to the demand word vector intention to obtain the second word vector feature of the conversation text, the embodiment of the present disclosure can process the second word vector feature to obtain the sentence vector feature of the conversation text.

[0108] In the disclosed embodiment, when processing the second word vector feature to obtain the sentence vector feature of the dialogue text, the weights of the second word vector feature can be calculated respectively, and then the second word vector feature can be weighted to obtain the weighted feature as the sentence vector feature of the dialogue text, and then the sentence vector feature can be processed by the intelligent dialogue system based on the relevant algorithms involved in artificial intelligence AI, and the sentence vector feature can be matched and retrieved by the intelligent dialogue system based on the AI ​​processing algorithm to obtain the response text corresponding to the dialogue text.

[0109] In this embodiment, by obtaining a dialogue text based on robot process automation RPA, determining a first word vector feature of the dialogue text based on artificial intelligence AI, determining a user's demand word vector intention, and processing the first word vector feature according to the demand word vector intention to obtain a second word vector feature of the dialogue text, RPA can be used in combination with artificial intelligence AI to pre-identify the user's demand word vector intention, so as to adjust and process the first word vector feature of the dialogue text, ensure the accuracy of word vector recognition and matching, effectively improve the accuracy of text feature generation, and improve the recall performance of the intelligent dialogue system. By determining the first word vector of the dialogue text based on artificial intelligence AI, determining a first word frequency corresponding to the first word vector based on a smoothed inverse word frequency algorithm, and determining a first word weight of the first word vector according to the first word frequency, the first word weight value of the first word vector can be determined. Since the first word weight can be used to perform weighted processing on the first word vector to obtain a sentence vector feature, the accuracy of the generated sentence vector feature can be improved, and a more accurate response text corresponding to the dialogue text can be generated to assist in generating.

[0110] Figure 3 It is a flowchart of a method for generating text features combining RPA and AI proposed in another embodiment of the present disclosure.

[0111] like Figure 3 As shown, the text feature generation method combining RPA and AI includes:

[0112] S301: Acquire the conversation text based on the Robotic Process Automation (RPA).

[0113] S302: Determine the first word vector feature of the conversation text based on artificial intelligence AI.

[0114] S303: Determine the user's demand word vector intention.

[0115] The description of S301 - S303 can be referred to the above embodiment for examples, and will not be repeated here.

[0116] S304: Process the required word vector intent based on AI to obtain multiple intent word vectors.

[0117] The intention word vector refers to the word vector obtained by mapping the user's demand word vector intention to the vector space. The demand word vector intention can be vectorized to obtain the intention word vector.

[0118] In the disclosed embodiment, after determining the user's demand word vector intention, the demand word vector intention can be processed based on artificial intelligence AI to obtain multiple intention word vectors, and the demand word vector intention can be encoded and vectorized using an encoding model to map the user's demand word vector intention to a vector space to obtain multiple word vectors output by the encoding model, and the multiple word vectors output by the encoding model can be used as multiple intention word vectors obtained by processing the demand word vector intention based on artificial intelligence AI.

[0119] S305: Perform principal component decomposition on the multiple intent word vectors to obtain common semantic information between the multiple intent word vectors.

[0120] Among them, principal component decomposition refers to converting multiple intent word vectors to obtain common semantic information between the multiple intent word vectors, so that the processed intent word vectors are unrelated intent word vectors.

[0121] After the above-mentioned AI-based processing of the demand word vector intent to obtain multiple intent word vectors, the embodiment of the present disclosure can perform principal component decomposition on the multiple intent word vectors to obtain common semantic information between the multiple intent word vectors.

[0122] In an embodiment of the present disclosure, when performing principal component decomposition on multiple intention word vectors to obtain common semantic information between the multiple intention word vectors, the intention word vectors can be constructed into a word vector matrix, the weighted average of the multiple intention word vectors can be calculated, the projection data information of the multiple intention word vectors in the word vector matrix can be calculated, and the calculated projection data information can be used as the common semantic information between the multiple intention vectors.

[0123] S306: Determine a target intent word vector from multiple intent word vectors based on common semantic information.

[0124] Among them, the target intent word vector refers to the irrelevant intent word vector obtained after removing some intent word vectors containing common semantic information from multiple intent word vectors. There is no correlation between the target intent word vectors.

[0125] In the disclosed embodiment, multiple intent word vectors can be transformed based on common semantic information, and some common intent word vectors can be removed from multiple intent word vectors containing relevant intent word vectors to obtain irrelevant intent word vectors as target intent word vectors.

[0126] In this embodiment, the intention of the demand word vector is processed based on AI to obtain multiple intention word vectors, and the multiple intention word vectors are decomposed into principal components to obtain the common semantic information between the multiple intention word vectors. According to the common semantic information, the target intention word vector is determined from the multiple intention word vectors, so that the common parts in the intention word vectors can be removed to obtain irrelevant target intention word vectors. Since the sentence vector features are generated by combining the target intention word vectors, the generation effect of the sentence vector features can be guaranteed, and the accuracy of the generated sentence vector features can be effectively improved, thereby helping to improve the recall performance of the intelligent dialogue system.

[0127] S307: If the target intent word vector matches the first word vector, the first word weight is increased to be the target word weight.

[0128] In the disclosed embodiment, after determining the target intention word vector from multiple intention word vectors based on the common semantic information as described above, the target intention word vector can be matched with the first word vector. If the target intention word vector and the first word vector match, the first word weight can be increased to be the target word weight. The first word weight corresponding to the first word vector can be increased to obtain the increased first word weight, and the increased first word weight can be used as the target word weight.

[0129] S308: Using the first word vector and the target word weight as the second word vector features.

[0130] In the disclosed embodiment, when the target intent word vector and the first word vector are matched, after the first word weight is increased to be the target word weight, the first word vector and the increased target word weight can be used as the second word vector features.

[0131] S309: If the target intention word vector matches the first word vector, the first word vector is described and processed based on the target word weight to obtain the sentence vector features of the dialogue text, wherein the intelligent dialogue system processes the sentence vector features based on AI to obtain the response text of the dialogue text.

[0132] In the disclosed embodiment, if the target intent word vector matches the first word vector, the first word weight is increased and processed as the target word weight, and the first word vector and the target word weight are used as the second word vector features. The first word vector can then be described based on the target word weight to obtain the sentence vector features of the conversation text.

[0133] In the disclosed embodiment, when the first word vector is described based on the target word weight, the first word vector can be weighted and described based on the target word weight obtained after the increase processing to obtain the sentence vector feature generated after the description processing, and the generated sentence vector feature is used as the sentence vector feature of the dialogue text, and then the intelligent dialogue system can be used to process the sentence vector feature based on artificial intelligence AI to obtain the response text corresponding to the dialogue text.

[0134] Optionally, in some embodiments, if the target intention word vector does not match the first word vector, the target intention word vector is used to update the first word vector, the requirement word vector intention is input into the weight parsing model in the AI ​​to obtain the intention word weight corresponding to the target intention word vector, the intention word weight is configured for the target intention word vector, and the target intention word vector and the intention word weight are used as the second word vector feature. If the target intention word vector does not match the first word vector, the target intention word vector is described and processed based on the intention word weight to obtain the sentence vector feature of the dialogue text, wherein the intelligent dialogue system processes the sentence vector feature based on AI to obtain the response text of the dialogue text, so that when the target intention word vector does not match the first word vector, the first word vector can be updated and the sentence vector feature of the dialogue text can be generated based on the intention word weight and the target intention word vector, thereby ensuring the comprehensiveness of the sentence vector feature processing of the dialogue text generated according to the first word vector, effectively improving the text feature generation processing effect, and assisting in improving the recall performance of the intelligent dialogue system.

[0135] Among them, the pre-trained weight parsing model can complete the task of weight parsing and configuring the vector, and can be used to configure the word weights of the word vector.

[0136] In the disclosed embodiment, the target intent word vector can be matched with the first word vector. If the target intent word vector does not match the first word vector, the target intent word vector is used to update the first word vector. Then, the requirement word vector intent can be input into a pre-trained weight parsing model, and the weight parsing model can be used to perform weight parsing on the requirement word vector intent to obtain the output result of the pre-trained weight parsing model. The output result of the pre-trained weight parsing model is used as the intent word weight. Then, the target intent word vector can be configured with the intent word weight, and the target intent word vector and the intent word weight can be used as the second word vector features.

[0137] In the disclosed embodiment, if the target intent word vector does not match the first word vector, the target intent word vector can be weighted described based on the intent word weight to obtain the vector feature after the weighted description as the sentence vector feature of the dialogue text, and then the intelligent dialogue system can be used to process the sentence vector feature based on artificial intelligence AI to obtain the response text of the dialogue text.

[0138] In this embodiment, by obtaining a conversation text based on robot process automation RPA, determining a first word vector feature of the conversation text based on artificial intelligence AI, determining the user's demand word vector intention, and processing the first word vector feature according to the demand word vector intention to obtain a second word vector feature of the conversation text, RPA can be used in combination with artificial intelligence AI to pre-identify the user's demand word vector intention, so as to adjust and process the first word vector feature of the conversation text, ensure the accuracy of word vector recognition and matching, improve the recall performance of the intelligent dialogue system, and effectively improve the accuracy of text feature generation. By processing the demand word vector intention based on AI to obtain multiple intention word vectors, the multiple intention word vectors are principally decomposed to obtain common semantic information between the multiple intention word vectors, and the target intention word vector is determined from the multiple intention word vectors based on the common semantic information, so that the common part in the intention word vector can be removed to obtain an irrelevant target intention word vector. Since the sentence vector feature is generated by combining the target intention word vector, the generation effect of the sentence vector feature can be guaranteed, the accuracy of the generated sentence vector feature is effectively improved, and the recall performance of the intelligent dialogue system is assisted in improving.

[0139] Figure 4 It is a structural diagram of a text feature generation device combining RPA and AI proposed in one embodiment of the present disclosure.

[0140] like Figure 4 As shown, the text feature generation device 40 combining RPA and AI is applied to an intelligent dialogue system supporting natural language processing NLP. The intelligent dialogue system can be registered and used by users, including:

[0141] An acquisition module 401 is used to acquire a conversation text based on a robotic process automation (RPA);

[0142] A first determination module 402, configured to determine a first word vector feature of the conversation text based on artificial intelligence AI;

[0143] The second determination module 403 is used to determine the user's demand word vector intention; and

[0144] The first processing module 404 is used to process the first word vector feature according to the required word vector intention to obtain the second word vector feature of the dialogue text.

[0145] In some embodiments of the present disclosure, Figure 5 As shown, Figure 5 : is a structural diagram of a text feature generation device combining RPA and AI proposed in another embodiment of the present disclosure, and also includes:

[0146] The second processing module 405 is used to process the second word vector feature to obtain the sentence vector feature of the dialogue text, wherein the intelligent dialogue system processes the sentence vector feature based on AI to obtain the response text of the dialogue text.

[0147] In some embodiments of the present disclosure, the first determining module 402 includes:

[0148] A first determination submodule 4021, configured to determine a first word vector of the dialogue text based on AI;

[0149] A second determining submodule 4022, configured to determine a first word frequency corresponding to the first word vector based on a smoothed inverse word frequency algorithm;

[0150] The third determination submodule 4023 is used to determine the first word weight of the first word vector according to the first word frequency, wherein the first word vector and the first word weight are jointly used as the first word vector feature.

[0151] In some embodiments of the present disclosure, it further includes:

[0152] The third processing module 406 is used to process the intention of the demand word vector based on AI after determining the intention of the user's demand word vector to obtain multiple intention word vectors;

[0153] The fourth processing module 407 is used to perform principal component decomposition on the multiple intention word vectors to obtain common semantic information between the multiple intention word vectors;

[0154] The third determination module 408 is used to determine a target intention word vector from multiple intention word vectors based on common semantic information.

[0155] In some embodiments of the present disclosure, the first processing module 404 is specifically configured to:

[0156] If the target intent word vector matches the first word vector, the first word weight is increased to be the target word weight;

[0157] The first word vector and the target word weight are used as the second word vector features.

[0158] In some embodiments of the present disclosure, the second processing module 405 is specifically configured to:

[0159] If the target intent word vector matches the first word vector, the first word vector is described based on the target word weight to obtain the sentence vector feature of the conversation text.

[0160] In some embodiments of the present disclosure, the first processing module 404 is further configured to:

[0161] If the target intent word vector does not match the first word vector, the target intent word vector is used to update the first word vector;

[0162] Configure the intent word weight for the target intent word vector;

[0163] The target intent word vector and the intent word weight are used as the second word vector features.

[0164] In some embodiments of the present disclosure, the second processing module 405 is further configured to:

[0165] If the target intent word vector does not match the first word vector, the target intent word vector is described based on the intent word weight to obtain the sentence vector feature of the conversation text.

[0166] In some embodiments of the present disclosure, the first processing module 404 is further configured to:

[0167] Before configuring the intent word weight for the target intent word vector, the requirement word vector intent is input into the weight parsing model in the AI ​​to obtain the intent word weight corresponding to the target intent word vector.

[0168] With the above Figures 1 to 3 Corresponding to the text feature generation method combining RPA and AI provided in the embodiment, the present disclosure also provides a text feature generation device combining RPA and AI. Figures 1 to 3 The text feature generation method combining RPA and AI provided in the embodiment corresponds to each other, so the implementation method of the text feature generation method combining RPA and AI is also applicable to the text feature generation device combining RPA and AI provided in the embodiment of the present disclosure, and will not be described in detail in the embodiment of the present disclosure.

[0169] The functions of each module in each device of the embodiments of the present disclosure can be found in the corresponding description of the above method, which will not be repeated here.

[0170] In this embodiment, by obtaining the conversation text based on robot process automation RPA, determining the first word vector feature of the conversation text based on artificial intelligence AI, determining the user's demand word vector intention, and processing the first word vector feature according to the demand word vector intention to obtain the second word vector feature of the conversation text, RPA can be combined with artificial intelligence AI to pre-identify the user's demand word vector intention, so as to adjust the first word vector feature of the conversation text, ensure the accuracy of word vector recognition and matching, effectively improve the accuracy of text feature generation, and improve the recall performance of the intelligent dialogue system.

[0171] In order to implement the above embodiments, the present disclosure also proposes an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the text feature generation method combining RPA and AI proposed in the above embodiments of the present disclosure is implemented.

[0172] Figure 6 FIG. 1 is a block diagram showing a structure of an electronic device according to an embodiment of the present disclosure. Figure 6 As shown, the electronic device 60 includes: a memory 610 and a processor 620, wherein the memory 610 stores a computer program that can be run on the processor 620. When the processor 620 executes the computer program, the text feature generation method combining RPA and AI in the above embodiment is implemented. The number of the memory 610 and the processor 620 can be one or more.

[0173] The electronic device 60 further includes:

[0174] The communication interface 630 is used to communicate with external devices and perform data exchange transmission.

[0175] If the memory 610, the processor 620 and the communication interface 630 are implemented independently, the memory 610, the processor 620 and the communication interface 630 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 6 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0176] Optionally, in a specific implementation, if the memory 610, the processor 620 and the communication interface 630 are integrated on a chip, the memory 610, the processor 620 and the communication interface 630 can communicate with each other through an internal interface.

[0177] An embodiment of the present disclosure provides a computer-readable storage medium storing a computer program, which implements the method provided in the embodiment of the present disclosure when the program is executed by a processor.

[0178] The embodiment of the present disclosure also provides a chip, which includes a processor for calling and executing instructions stored in the memory from the memory, so that a communication device equipped with the chip executes the method provided by the embodiment of the present disclosure.

[0179] The embodiment of the present disclosure also provides a chip, including: an input interface, an output interface, a processor and a memory. The input interface, the output interface, the processor and the memory are connected through an internal connection path. The processor is used to execute the code in the memory. When the code is executed, the processor is used to execute the method provided in the embodiment of the application.

[0180] It should be understood that the processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. It is worth noting that the processor may be a processor supporting the Advanced RISC Machines (ARM) architecture.

[0181] Further, optionally, the above-mentioned memory may include a read-only memory and a random access memory, and may also include a non-volatile random access memory. The memory may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may include a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may include a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available. For example, static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM) and direct memory bus random access memory (DR RAM).

[0182] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function according to the present disclosure is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium.

[0183] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present disclosure. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they are contradictory.

[0184] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of the present disclosure, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0185] Any process or method description described in the flowchart or otherwise herein can be understood to represent a module, segment or portion of code including one or more executable instructions for implementing the steps of a specific logical function or process. And the scope of the preferred embodiments of the present disclosure includes other implementations, in which the functions may not be performed in the order shown or discussed, including in a substantially simultaneous manner or in a reverse order according to the functions involved.

[0186] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, which can be embodied in any computer-readable medium for use by an instruction execution system, apparatus or device (such as a computer-based system, a system including a processor or other system that can fetch instructions from an instruction execution system, apparatus or device and execute instructions), or used in combination with these instruction execution systems, apparatuses or devices.

[0187] It should be understood that the various parts of the present disclosure may be implemented with hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods may be implemented with software or firmware stored in a memory and executed by a suitable instruction execution system. All or part of the steps of the above embodiment method may be completed by instructing the relevant hardware through a program, which may be stored in a computer-readable storage medium, and when the program is executed, includes one or a combination of the steps of the method embodiment.

[0188] In addition, each functional unit in each embodiment of the present disclosure may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the above-mentioned integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium. The storage medium may be a read-only memory, a disk or an optical disk, etc.

[0189] The above is only a specific implementation of the present disclosure, but the protection scope of the present disclosure is not limited thereto. Any technician familiar with the technical field can easily think of various changes or substitutions within the technical scope disclosed in the present disclosure, which should be included in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be based on the protection scope of the claims.

Claims

1. A text feature generation method combining RPA and AI, characterized in that: Applied to an intelligent dialogue system supporting natural language processing (NLP), the intelligent dialogue system is available for user registration and use, and the method includes: Get the conversation text based on Robotic Process Automation (RPA); Determine a first word vector feature of the conversation text based on artificial intelligence AI; Determining the user's demand word vector intention, wherein the user's demand word vector intention refers to a word feature intention pre-set according to the user's customized semantic association between words, and the demand word vector is used to perform customized zooming processing on the word vector of the conversation text; and Processing the first word vector feature according to the demand word vector intention to obtain a second word vector feature of the conversation text; After determining the user's demand word vector intention, the method further includes: Processing the demand word vector intent based on the AI ​​to obtain multiple intent word vectors; Performing principal component decomposition on the multiple intention word vectors to obtain common semantic information between the multiple intention word vectors; Determining a target intention word vector from the multiple intention word vectors according to the common semantic information; The processing of the first word vector feature according to the demand word vector intention to obtain a second word vector feature of the conversation text includes: If the target intent word vector matches the first word vector, the first word weight is increased to be the target word weight; Using the first word vector and the target word weight as the second word vector features; The processing of the first word vector feature according to the demand word vector intention to obtain a second word vector feature of the conversation text includes: If the target intention word vector does not match the first word vector, updating the first word vector using the target intention word vector; Configuring an intent word weight for the target intent word vector; The target intent word vector and the intent word weight are used as the second word vector features.

2. The method according to claim 1, characterized in that Also includes: The second word vector feature is processed to obtain a sentence vector feature of the dialogue text, wherein the intelligent dialogue system processes the sentence vector feature based on the AI ​​to obtain a response text of the dialogue text.

3. The method according to claim 2, characterized in that The determining the first word vector feature of the conversation text based on artificial intelligence AI includes: Determine a first word vector of the conversation text based on the AI; Determine a first word frequency corresponding to the first word vector based on a smoothed inverse word frequency algorithm; A first word weight of the first word vector is determined based on the first word frequency, wherein the first word vector and the first word weight are jointly used as features of the first word vector.

4. The method according to claim 1, characterized in that The processing of the second word vector feature to obtain a sentence vector feature of the conversation text includes: If the target intention word vector matches the first word vector, the first word vector is described and processed based on the target word weight to obtain the sentence vector feature of the conversation text.

5. The method according to claim 1, characterized in that The processing of the second word vector feature to obtain a sentence vector feature of the conversation text includes: If the target intent word vector does not match the first word vector, the target intent word vector is described and processed based on the intent word weight to obtain a sentence vector feature of the conversation text.

6. The method according to claim 1, characterized in that Before configuring the target intention word vector with an intention word weight, the method further includes: The demand word vector intention is input into the weight parsing model in the AI ​​to obtain the intention word weight corresponding to the target intention word vector.

7. A text feature generation device combining RPA and AI, characterized in that: An intelligent dialogue system is applied to support natural language processing (NLP), and the intelligent dialogue system can be registered and used by users. The device includes: The acquisition module is used to obtain the conversation text based on the Robotic Process Automation (RPA); A first determination module, used to determine a first word vector feature of the dialogue text based on artificial intelligence AI; A second determination module is used to determine the user's demand word vector intention, wherein the user's demand word vector intention refers to a word feature intention pre-set according to the user's customized semantic association between words, and the demand word vector is used to perform customized zoom processing on the word vector of the dialogue text; and A first processing module, configured to process the first word vector feature according to the demand word vector intention to obtain a second word vector feature of the dialogue text; Also includes: A third processing module is used for processing the demand word vector intention based on the AI ​​after determining the demand word vector intention of the user to obtain multiple intention word vectors; A fourth processing module, configured to perform principal component decomposition on the plurality of intention word vectors to obtain common semantic information between the plurality of intention word vectors; A third determination module, configured to determine a target intention word vector from the plurality of intention word vectors according to the common semantic information; The first processing module is specifically used for: If the target intent word vector matches the first word vector, the first word weight is increased to be the target word weight; Using the first word vector and the target word weight as the second word vector features; The first processing module is further used for: If the target intention word vector does not match the first word vector, updating the first word vector using the target intention word vector; Configuring an intent word weight for the target intent word vector; The target intent word vector and the intent word weight are used as the second word vector features.

8. The device according to claim 7, characterized in that Also includes: The second processing module is used to process the second word vector feature to obtain the sentence vector feature of the dialogue text, wherein the intelligent dialogue system processes the sentence vector feature based on the AI ​​to obtain the response text of the dialogue text.

9. The device according to claim 8, characterized in that The first determining module includes: A first determination submodule, configured to determine a first word vector of the dialogue text based on the AI; A second determination submodule, configured to determine a first word frequency corresponding to the first word vector based on a smoothed inverse word frequency algorithm; The third determination submodule is used to determine the first word weight of the first word vector according to the first word frequency, wherein the first word vector and the first word weight are jointly used as the first word vector feature.

10. The device according to claim 8, characterized in that The second processing module is specifically used for: If the target intention word vector matches the first word vector, the first word vector is described and processed based on the target word weight to obtain the sentence vector feature of the conversation text.

11. The device according to claim 8, characterized in that The second processing module is further used for: If the target intent word vector does not match the first word vector, the target intent word vector is described and processed based on the intent word weight to obtain a sentence vector feature of the conversation text.

12. The device according to claim 7, characterized in that The first processing module is further used for: Before configuring the intention word weight for the target intention word vector, the requirement word vector intention is input into the weight parsing model in the AI ​​to obtain the intention word weight corresponding to the target intention word vector.

13. An electronic device, comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the text feature generation method combining RPA and AI as described in any one of claims 1-6.

14. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to enable the computer to execute the text feature generation method combining RPA and AI as described in any one of claims 1-6.

Images