Auxiliary file reading method and device, computer equipment and storage medium

Abstract

The invention relates to the technical field of voice semantics and financial science and technology, and discloses a file auxiliary reading method and device, computer equipment and a storage medium, and the method comprises the steps: recognizing a target user according to user information; performing voice conversion on the user protocol text of the target user to obtain protocol voice; collecting user interaction voice corresponding to the protocol voice, and identifying an interaction voice text of the user interaction voice; calculating protocol explanation content according to the interactive voice text; and collecting interaction data according to the protocol explanation content to obtain auxiliary reading data of the target user. By performing voice conversion on the user protocol text of the target user and collecting the interaction data, the problems proposed by the target user can be solved in a targeted manner, meanwhile, the service dispute and the service conflict are avoided, and the auxiliary reading effect of the protocol text is effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of speech semantics and financial technology, and in particular to a document auxiliary reading method, device, computer equipment and storage medium. Background Art

[0002] With the continuous advancement of Internet technology and the aging of the population, the convenience and accessibility of online financial services have become the focus of financial institutions. However, when older people use smart devices such as smartphones or tablets to handle online business, they may encounter problems with vision and comprehension, which often leads to conversion losses or subsequent business disputes.

[0003] In the prior art, more and more online financial business transactions require users to read personal insurance agreements or sign online contracts, including but not limited to logging into the system, purchasing goods and services, etc. For example, when it comes to such agreement documents, conventional smartphones or tablets require users to click on a text link and browse through the multi-page lengthy contract agreement in the newly opened page. After reading, users usually need to tick the box to agree or sign an additional handwritten signature. After completing the above steps, they can enter the subsequent login, payment and other processes. The system will record the user's tick and handwritten signature actions to facilitate evidence when business disputes arise. However, the terms of the agreement are numerous and complex, making it difficult for the elderly to understand. The screen of smart mobile devices is small, the operating space is limited, and it is easy to make mistakes. At the same time, they do not adopt a method suitable for the elderly to read. Therefore, how to improve the effect of assisted reading of online agreement documents has become an urgent problem to be solved. Summary of the invention

[0004] The present invention provides a document auxiliary reading method, device, computer equipment and storage medium to solve the technical problem that the agreement document has many clauses and the smart device is easy to misoperate, resulting in difficulty in reading the agreement document.

[0005] In a first aspect, a file auxiliary reading method is provided, comprising:

[0006] Acquire user information, and identify a target user according to the user information;

[0007] Performing voice conversion on the user agreement text of the target user to obtain the agreement voice;

[0008] Collecting the user interaction voice corresponding to the protocol voice, and identifying the interactive voice text of the user interaction voice;

[0009] Calculate the protocol explanation content of the user interactive speech according to the interactive speech text;

[0010] The interactive data is collected according to the protocol to explain the content and obtain auxiliary reading data of the target user.

[0011] In a second aspect, a file auxiliary reading device is provided, comprising:

[0012] A target user identification module, used to obtain user information and identify the target user according to the user information;

[0013] A voice conversion module, used for performing voice conversion on the user agreement text of the target user to obtain the agreement voice;

[0014] An interactive speech recognition module, used to collect the user interactive speech corresponding to the protocol speech, and recognize the interactive speech text of the user interactive speech;

[0015] A protocol explanation content generation module, used for calculating the protocol explanation content of the user interactive speech according to the interactive speech text;

[0016] The auxiliary reading data collection module is used to collect interactive data according to the protocol explanation content to obtain the auxiliary reading data of the target user.

[0017] In a third aspect, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned file assisted reading method when executing the computer program.

[0018] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned file assisted reading method are implemented.

[0019] In the scheme implemented by the above-mentioned document auxiliary reading method, device, computer equipment and storage medium, user information can be received through the client, and the target user can be identified according to the user information; the user agreement text of the target user can be converted into speech to obtain the agreement speech; the user interaction speech corresponding to the agreement speech is collected, and the interactive speech text of the user interaction speech is identified; the agreement explanation content is calculated according to the interactive speech text; the interactive data is collected according to the agreement explanation content to obtain the auxiliary reading data of the target user, and the auxiliary reading data is fed back to the client. In the present invention, for complex insurance entities such as insurance business, by identifying the target user, the target user who needs to be assisted in reading can be targeted to help them better understand and improve the effect of protocol document reading; the user agreement text of the target user is converted into speech, and the user agreement text is voice-read in combination with the protocol speech to improve the effect of assisted reading of the protocol text; the user interaction speech is collected to obtain the possible problems of the target user in real time, and the protocol explanation content of the user interaction speech can solve the problems raised by the target user in a targeted manner, effectively improving the effect of assisted reading; the interactive data is collected according to the agreement explanation content to obtain the auxiliary reading data of the target user, which can be used for evidence in the future to avoid business disputes and business conflicts, and further improve the effect of assisted reading of protocol documents. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0021] Figure 1 is a schematic diagram of an application environment of a file auxiliary reading method in one embodiment of the present invention;

[0022] Figure 2 is a schematic diagram of a flow chart of a method for assisting file reading in one embodiment of the present invention;

[0023] Figure 3 It is a schematic diagram of a flow chart of identifying interactive voice text of user interactive voice provided by an embodiment of the present invention;

[0024] Figure 4 is a structural schematic diagram of a document auxiliary reading device in one embodiment of the present invention;

[0025] Figure 5 is a schematic diagram of a structure of a computer device in one embodiment of the present invention;

[0026] Figure 6It is another structural schematic diagram of a computer device in one embodiment of the present invention. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] The file auxiliary reading method provided by the embodiment of the present invention can be applied in the following aspects: Figure 1 In the application environment, the client communicates with the server through the network. The server can receive user information through the client, identify the target user according to the user information; perform voice conversion on the user agreement text of the target user to obtain the agreement voice; collect the user interaction voice corresponding to the agreement voice, identify the interactive voice text of the user interaction voice; calculate the agreement explanation content according to the interactive voice text; collect the interactive data according to the agreement explanation content, obtain the auxiliary reading data of the target user, and feed the auxiliary reading data back to the client. In the present invention, for complex insurance entities such as insurance business, the target user is identified, and the target user who needs to be assisted in reading can be targeted to help them better understand and improve the effect of protocol document reading; perform voice conversion on the user agreement text of the target user, and perform voice reading of the user agreement text in combination with the agreement voice to improve the effect of assisted reading of the agreement text; collect the user interaction voice, obtain the possible problems of the target user in real time, and the agreement explanation content of the user interaction voice can solve the problems raised by the target user in a targeted manner, effectively improving the effect of assisted reading; collect the interactive data according to the agreement explanation content, obtain the auxiliary reading data of the target user, and provide evidence in the future to avoid business disputes and business conflicts, and further improve the effect of assisted reading of the agreement document. The client may be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, and portable wearable devices. The server may be implemented by an independent server or a server cluster consisting of multiple servers. The present invention is described in detail below through specific embodiments.

[0029] See also Figure 2 As shown, Figure 2 A flowchart of a file auxiliary reading method provided by an embodiment of the present invention includes the following steps:

[0030] S1. Obtain user information, and identify a target user based on the user information.

[0031] In an embodiment of the present invention, the user information may be information about a smart device used by the user, for example, the model of the user's smart device, whether a specific system has acquired an elderly mode turned on, an image of the user taken on the smart device, etc.

[0032] Specifically, the target users refer to elderly users who need to assist in reading agreement documents. As an important user group of financial services, the needs of elderly users include savings deposits, insurance needs, payment settlement, investment and financing needs, and other personalized financial services. With the increase in the number of elderly people, more and more social assets are gathered in the hands of the elderly, stimulating the growth of demand for financial services. Therefore, targeting elderly users can help the elderly see the text on the screen of smart devices more clearly and improve the comfort of reading. At the same time, it strengthens the risk warnings and information verification of key businesses and key links, significantly improving the convenience, safety and satisfaction of the elderly in using smart devices for financial business, and helping them better integrate into digital financial life.

[0033] In detail, an agreement document refers to a formal document drawn up between two or more parties to regulate a certain behavior or transaction and clarify the rights, obligations and responsibilities of each party. An agreement document usually contains the terms and conditions of the agreement, as well as the consensus and commitment of all parties involved. An agreement document is legally binding. Once signed, all parties must abide by the terms therein, otherwise they may face legal consequences. The terms in the agreement document need to be clear and unambiguous to avoid differences in interpretation. At the same time, the agreement document contains all necessary details, such as information about the parties, the term of the agreement, fees, limitations of liability, confidentiality clauses, etc. For example, an agreement document can be a commercial contract involving a transaction of goods or services between a buyer and a seller, including terms such as price, quantity, and delivery time; a lease agreement that stipulates the rights and obligations of the lessor and the lessee in terms of leased property (such as houses, equipment, etc.); a confidentiality agreement that stipulates the obligations of the parties in the process of exchanging confidential information and prevents information leakage, and financial agreements involving financial transactions, such as loan agreements, investment agreements, insurance contracts, etc. Agreement documents are important tools in commercial and legal activities, which can help ensure the transparency and fairness of financial business transactions and provide a basis for resolving possible disputes. Therefore, auxiliary reading is needed to better help target users understand the contents of the agreement document and ensure the interests of target users.

[0034] Among them, smart devices refer to portable mobile devices with powerful computing capabilities, Internet connection functions, multi-functional operating systems, and the ability to install and run various applications. For example, smart phones have touch screen interfaces that can install and run various applications, and tablets have larger screens than smart phones, which are suitable for media consumption and games and can also install a variety of applications.

[0035] In the embodiment of the present invention, the step of identifying the target user according to the user information includes:

[0036] Classifying the user information to obtain information to be identified for identifying user categories;

[0037] Perform information identification based on the information to be identified to obtain user category information;

[0038] The target user is determined according to the user category information.

[0039] In the embodiment of the present invention, the information to be identified refers to information that can indicate the category of the user, for example, the model of the smart device used, the usage mode of the smart device, and the image of the user.

[0040] Furthermore, information recognition is to identify whether each category of information is in the user category information corresponding to the corresponding elderly user. For example, the model of the smart device belongs to the preset list of commonly used smart device models for the elderly. The smart device specific system obtains the elderly mode that has been turned on, or the front camera of the smart device is used to capture a 1-frame user image, and the locally set large visual model is used to determine whether the user corresponding to the user image is an elderly person.

[0041] Exemplarily, the visual big model can be a visual language model (VLM, Visual Language Model), which combines image and natural language processing technology, and can generate descriptive text, visual question and answer, and image retrieval for images. The working principle of VLM involves an image encoder, a text encoder, a multimodal fusion module, and a generator. The image encoder converts the image into a high-dimensional feature vector, the text encoder converts the text into a corresponding semantic representation, and the multimodal fusion module is responsible for aligning and fusing the two representations to generate a unified cross-modal representation. Finally, the generator generates corresponding text output or image output based on the fused representation. Therefore, the visual big model can be used to identify whether the collected user image is the target user.

[0042] In detail, when there is at least one user category information corresponding to the elderly in the user category information, the user corresponding to the user category information can be determined as the target user, that is, the elderly user who needs assisted reading of the online protocol file.

[0043] In the embodiment of the present invention, by identifying the target user, the elderly users who need assisted reading can be specifically identified, and then when the elderly users read the agreement, they can be helped to better understand and improve the effect of reading the agreement document.

[0044] S2. Perform voice conversion on the user agreement text of the target user to obtain the agreement voice.

[0045] In an embodiment of the present invention, the user agreement text is a financial business text read by the target user. It is a formal written document involving the rights and obligations between the parties in the financial business. For example, it is an agreement signed between financial institutions or between financial institutions and customers to establish a cooperative relationship, which includes terms such as cooperation principles, rights and obligations of both parties, cooperation content, and dispute resolution methods.

[0046] In detail, speech conversion converts the text in the user agreement text into speech through the text-to-speech (TTS) technology, and converts the electronic text information in the user agreement text into audible speech output.

[0047] In the embodiment of the present invention, the voice conversion of the user agreement text of the target user to obtain the agreement voice includes:

[0048] The user agreement text is divided into text characters to obtain text characters:

[0049] Performing character feature encoding on the text characters to obtain a character code corresponding to each of the text characters;

[0050] Calculate the Mel spectrum corresponding to each of the text characters according to the character encoding;

[0051] The Mel spectrum is converted into speech to obtain the protocol speech corresponding to the user protocol text.

[0052] In the embodiment of the present invention, text segmentation is to decompose the complete text in the user agreement text into individual character units, and these characters can be letters, numbers, punctuation marks, spaces or any other text symbols.

[0053] In detail, the user agreement text can be divided into text characters through a preset natural language processing tool, for example, the nltk library and the spaCy library in Python. In Python, the string itself is iterable and can be directly used in a for loop. In JavaScript and Java, specific methods (such as a for...of loop or a toCharArray() method) are required to traverse each character in the user agreement text to obtain the text characters.

[0054] Furthermore, character feature encoding is to convert continuous numerical attributes into discrete numerical attributes. One-Hot encoding can be used to encode the character features of text characters to obtain character encoding. Among them, One-Hot encoding (also known as One-Hot encoding) is a special binary encoding form, which represents each value of a discrete feature as an independent binary bit, and only this bit is 1, and the rest of the bits are 0. For example, for a discrete feature with a value range of {0,1,2}, after One-Hot encoding, it may be represented as [1,0,0], [0,1,0], [0,0,1], corresponding to the original values ​​of 0, 1, and 2, respectively. Through One-Hot encoding, discrete features can be converted into binary form, eliminating the partial order relationship between features, so that the algorithm can treat each feature value more fairly, facilitate calculation, and improve the efficiency of subsequent calculations.

[0055] Furthermore, the Mel spectrum is a spectrum diagram that shows the distribution of the signal's energy in time and Mel-scale frequency. The Mel scale is a nonlinear frequency scale based on human auditory perception. By using the Mel spectrum as a representation of sound features, the human ear can more easily perceive changes in sound frequency.

[0056] In the embodiment of the present invention, text characters can be converted into frames in an audio signal through character encoding, and then the Mel spectrum corresponding to each text character can be extracted.

[0057] Specifically, calculating the Mel spectrum corresponding to each character in the text according to the character encoding includes:

[0058] Performing short-time Fourier transform on the character code to obtain power spectrum density;

[0059] Calculate the Mel filter spectrum of the character code according to the power spectrum density;

[0060] Logarithmically compress the Mel filter spectrum to obtain Mel spectrum corresponding to the characters in the text.

[0061] In an embodiment of the present invention, a short-time Fourier transform (STFT) is a method for analyzing the time-varying characteristics of a signal frequency. First, a window function (such as a Hanning window, a Hamming window, etc.) is selected and applied to each frame of the signal. The window function helps to reduce the disconnection at the frame boundary; the signal is divided into continuous frames, the length of each frame is usually fixed, and there is a certain overlap between frames to reduce information loss; each frame processed by the window function is subjected to a fast Fourier transform (FFT) to obtain the frequency content of each frame; the FFT results of all frames are combined to form a time-frequency representation. The horizontal axis represents time, the vertical axis represents frequency, and the color or brightness represents the energy at different times and frequencies to obtain a frequency domain representation of the character code.

[0062] Furthermore, the Mel filter spectrum is obtained by multiplying the power spectrum density obtained by the short-time Fourier transform with the preset Mel filter bank to obtain the Mel spectrum. This process involves multiplying the response of each filter with the corresponding short-time Fourier transform power spectrum, and then summing the results to obtain the energy value of each Mel filter.

[0063] In detail, logarithmic compression can simulate the nonlinear perception of sound intensity by the human ear, which usually involves taking the logarithm of 10 and then multiplying it by 20, that is, 20*log10(filter_banks), to obtain a protocol speech that is better perceived by the human ear.

[0064] In detail, the human ear's perception of sound is nonlinear. Logarithmic compression can make the energy value distribution of the Mel spectrum more uniform, reduce the dynamic range, and make the difference between low energy values ​​(low loudness) and high energy values ​​(high loudness) that are not easily perceived by the human ear be numerically amplified, thereby making subsequent processing and analysis more effective. Therefore, logarithmic compression can simulate the human ear's perception of the loudness of sound, better simulate the auditory characteristics of the human ear, reduce the dynamic range of the data, and improve the perceptibility of small signals, so that the Mel spectrum can be more accurately converted to speech and obtain the agreed speech.

[0065] Furthermore, the Mel spectrum corresponding to the text characters of the pre-trained seq2seq (sequence to sequence) model can be used, and the spectrum features can then be converted into an audible speech waveform through a vocoder to obtain the protocol speech.

[0066] In detail, the Seq2Seq (sequence to sequence) model is a deep learning architecture specifically designed to handle the mapping task from an input sequence to an output sequence. The Seq2Seq model consists of two main parts: the encoder and the decoder. The main responsibility of the encoder is to extract key features from the input sequence and effectively compress these features into a fixed-dimensional vector (also called a context vector or hidden state vector) or a series of hidden states. The task of the decoder is to gradually generate the target sequence based on the context vector or hidden state sequence provided by the encoder. The decoder is usually a generative model that predicts the next output unit (such as a word or character) at each time step until the complete output sequence is generated. The core idea of ​​the Seq2Seq model is to use the encoder network to encode the input sequence (such as a source language sentence) into a fixed-dimensional vector or a series of hidden states, and then use the decoder network to generate the target sequence (such as the target language sentence) word by word from this vector or hidden state. The whole process does not require manual design of complex language rules or intermediate representations, but allows the neural network to learn how to perform effective sequence conversion. The Seq2Seq model can output the protocol speech corresponding to the Mel spectrum.

[0067] Preferably, after determining the user agreement text, the text of the user agreement text can be enlarged on the target user's device interface, for example, the text of the user agreement text can be enlarged to 3 times, 4 times, etc., which is more convenient for the target user to read. At the same time, combined with the agreement voice, the effect of assisting reading of the agreement text is effectively improved.

[0068] S3. Collect the user interaction voice corresponding to the protocol voice, and identify the interactive voice text of the user interaction voice.

[0069] In the embodiment of the present invention, the user interaction voice may be the voice of the user asking questions, for example, "What does this clause mean?", "This clause seems a bit vague, can you explain it in detail?", "How long is the validity period of this clause?", etc. This can help users better understand the content of the agreement and ensure that their questions are answered.

[0070] Specifically, user voice can be collected in real time. When the user asks a question, the user interaction voice is compared with the read agreement voice to determine whether they are consistent, and then the read user agreement text content is adjusted in real time according to the user interaction voice.

[0071] In the embodiment of the present invention, refer to Figure 3 As shown, the interactive voice text of the user interactive voice is identified, including:

[0072] S31, performing voice pre-emphasis and voice noise reduction processing on the user interaction voice to obtain an enhanced voice signal;

[0073] S32, performing signal division on the enhanced speech signal to obtain a speech signal sequence;

[0074] S33, extracting voice features of the user interaction voice according to the voice signal sequence;

[0075] S34: Decode the voice feature signal to obtain the interactive voice text of the user interactive voice.

[0076] In an embodiment of the present invention, speech pre-emphasis and speech noise reduction processing are pre-processing of user interaction speech. Among them, pre-emphasis is to enhance the high-frequency component of the speech signal to compensate for the large attenuation of the high-frequency signal during the transmission process. In the speech signal, the high-frequency component (i.e., consonant) contains more information, while the frequency of vowels is generally low. Pre-emphasis can keep the low-frequency part of the user interaction speech unchanged and enhance the high-frequency part of the user interaction speech. Pre-emphasis can be achieved by a first-order high-pass filter. Signal denoising can be achieved by spectral subtraction or Wiener filtering. Through pre-emphasis and speech noise reduction processing, the useful information in the speech signal can be retained as much as possible, while removing noise and other interference, to obtain a more accurate enhanced speech signal.

[0077] In the embodiment of the present invention, signal division is to perform frame processing on the enhanced speech signal, that is, to divide the audio signal into small time intervals to obtain a speech signal sequence.

[0078] Furthermore, speech features are characteristic information in enhanced speech signals, such as Mel-frequency cepstral coefficients, linear prediction cepstral coefficients, etc., and speech features can be further extracted by applying a window function, such as a Hanning window, to each frame of the signal.

[0079] In detail, applying a window function to each frame of signal is an important step in digital signal processing, and its main purpose is to reduce spectrum leakage and improve the accuracy of spectrum analysis. Commonly used window functions include rectangular window, Hanning window, Hamming window and Blackman window. Each window function has its own specific spectral characteristics, such as main lobe width and sidelobe suppression level. These characteristics determine the applicability of the window function in different applications. Programming languages ​​(such as Python or MATLAB) can be used to generate window functions. For example, in MATLAB, built-in functions such as rectwin, hanning, hamming and blackman can be used to generate corresponding window functions, and the generated window function can be multiplied with the signal to achieve the application of window functions to each frame of signal.

[0080] Furthermore, the pre-trained acoustic model and language model can be used to convert speech to text, that is, mapping speech features to corresponding feature spaces, determining the text probability corresponding to each speech signal, combining the predicted texts, and obtaining the interactive speech text of the user's interactive speech.

[0081] In the embodiment of the present invention, by collecting user interactive voice in real time, possible problems of the target user can be obtained in real time and solved in a targeted manner, thereby effectively improving the effect of auxiliary reading.

[0082] S4. Calculate the protocol explanation content of the user interactive voice according to the interactive voice text.

[0083] In the embodiment of the present invention, the protocol explanation content provides clear and detailed explanations and guidance for the questions raised by the user in the interactive voice text, which can help the target user understand specific concepts, operation processes, product functions or solutions to problems. For example, for the interactive voice text "What does this other scope of legal permission mean?", a corresponding explanation is provided, explanation content is generated, and the user's question is explained.

[0084] In the embodiment of the present invention, the step of calculating the protocol explanation content of the user interactive speech according to the interactive speech text includes:

[0085] Determine the real-time protocol text corresponding to the protocol voice, compare the protocol text with the interactive voice text for consistency, and obtain a comparison result;

[0086] When the comparison result is consistent, searching the preset knowledge base for the protocol explanation content of the user interaction voice;

[0087] When the comparison result is inconsistent, the protocol explanation content of the user interaction voice is retrieved from the user protocol text.

[0088] In the embodiment of the present invention, the real-time protocol text is the content of the user protocol text being read aloud when the user interactive voice is collected, and the real-time protocol text is compared with the interactive voice text for consistency, wherein the consistency comparison can be performed based on the similarity between the real-time protocol text and the interactive voice text. When the similarity is greater than a preset similarity threshold, the comparison result is determined to be consistent, otherwise it is inconsistent.

[0089] In detail, the length of the longest common subtext of the real-time protocol text and the interactive voice text can be calculated according to the longest common subsequence algorithm to characterize their similarity, or by setting a sliding window of size N, the text content of the real-time protocol text and the interactive voice text is subjected to a window sliding operation in the form of a character stream or a word stream to form multiple text fragments of length N, each of which is called an N-tuple. Then, the ratio of the number of common N-tuples in the given two texts to the total number of N-tuples is calculated to characterize the similarity between the real-time protocol text and the interactive voice text. The similarity can also be calculated using the TF-IDF weight calculation method and the cosine similarity algorithm, and the consistency comparison result of the real-time protocol text and the interactive voice text is reflected by the similarity.

[0090] In detail, when the comparison result is consistent, it means that the target user may have questions about the content in the user agreement text and needs answers. Therefore, the protocol explanation content can be matched in the preset knowledge base to obtain the protocol explanation content, where the knowledge base is a pre-built database containing multiple protocol explanation contents. Relevant information can be retrieved from a large amount of data through RAG (Retrieval-Augmented Generation) technology, and then the final protocol explanation content is generated using a language model.

[0091] Furthermore, when the comparison result is inconsistent, it means that the user agreement text being read aloud may not be the content required by the target user. Therefore, the corresponding content is retrieved from the user agreement text as the agreement explanation content. For example, if the interactive voice text is "What does this other scope permitted by law mean?", the word splitting algorithm can be used to calculate the match from long to short, and finally the three "scope permitted by law" locations in the user agreement text are matched, and then jump to the location in turn to obtain the agreement explanation content.

[0092] Furthermore, after obtaining the protocol explanation content, the corresponding voice of the protocol explanation content can be generated for reading aloud. At the same time, after the protocol explanation content is read aloud, the content corresponding to the last corresponding real-time protocol text is returned for voice reading until the user confirms that the user protocol text assisted reading is completed.

[0093] In the embodiment of the present invention, the protocol explanation content can provide clear and detailed explanations and guidance for the user's real-time questions, helping the target user to better understand the user agreement text.

[0094] S5. Explain the content and collect interactive data according to the protocol to obtain auxiliary reading data of the target user.

[0095] In the embodiment of the present invention, the interaction data is archived for all voice and interaction operations during the user consultation and approval interaction process when the target user is assisting in reading the user agreement text, and can be used as evidence when a business dispute occurs.

[0096] In the embodiment of the present invention, the step of collecting interactive data according to the protocol to explain the content and obtain auxiliary reading data of the target user includes:

[0097] Generate explanation voice according to the explanation content of the protocol;

[0098] Invoking a preset user interaction component according to the explanation voice, and collecting component interaction data of the user interaction component;

[0099] The component interaction data, the interaction voice and the protocol explanation content are collected to obtain auxiliary reading data of the target user.

[0100] In the embodiment of the present invention, the explanation voice is a voice signal corresponding to the protocol explanation content, and the user protocol text is voice read according to the voice signal until the content in the user protocol text is read aloud, and a preset user interaction component is called.

[0101] In detail, the user interaction component can be a check box, signature, or other component that requires the target user to agree to end the assisted reading and is displayed in full screen on the target user's smart device, or a component that voice-inquires the target user whether he agrees to end the assisted reading, to confirm the end of the assisted reading of the agreement document.

[0102] Furthermore, user interaction components are elements used to implement interaction between the target user and the system. These components are the basis of the user interface (UI), which allows the target user to perform operations, receive feedback, and communicate with the system to obtain component interaction data.

[0103] In detail, the user interaction component is the basic unit that constitutes the system or software. It has specific functions and interfaces and can be developed, tested and replaced independently of other parts of the system. The user interaction component can be a user interface component, such as buttons, text boxes, lists, etc., which are the direct interface for users to interact with the software system; a business logic component that contains code to implement specific business rules and logic, such as order processing, user authentication, etc.; a data access component that is responsible for interacting with a database or other data storage to perform data addition, deletion, modification and query operations; a component that provides specific services, such as a service component for mail services, payment gateway interfaces, etc.; a hardware abstraction component that provides an abstract interface to hardware devices, such as a graphics processing unit (GPU), a sensor, etc.

[0104] In an embodiment of the present invention, all voice and interactive operations of the target user during the interactive process of consulting when assisting in reading the user agreement text are collected and archived, and the assisted reading data of the target user during the assisted reading is obtained, which can be used as evidence in future business disputes to avoid business disputes and business conflicts, and further improve the effect of assisted reading of agreement documents.

[0105] It can be seen that in the above scheme, for complex insurance entities under insurance business, target users can be identified, and target users who need assisted reading can be specifically identified to help them better understand and improve the effect of reading the agreement document; the user agreement text of the target user is converted into voice, and the user agreement text is voice-read in combination with the agreement voice to improve the effect of assisted reading of the agreement text; user interaction voice is collected to obtain possible problems of the target user in real time, and the agreement explanation content of the user interaction voice can solve the problems raised by the target user in a targeted manner, effectively improving the effect of assisted reading; interaction data is collected according to the agreement explanation content to obtain the assisted reading data of the target user, which can be used for evidence in the future to avoid business disputes and business conflicts, and further improve the effect of assisted reading of agreement documents.

[0106] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0107] In one embodiment, a document assisted reading device is provided, and the document assisted reading device corresponds one-to-one to the document assisted reading method in the above embodiment. Figure 4 As shown, the document auxiliary reading device 400 includes a target user identification module 401, a voice conversion module 402, an interactive voice recognition module 403, a protocol explanation content generation module 404 and an auxiliary reading data collection module 405. The functional modules are described in detail as follows:

[0108] The target user identification module 401 is used to obtain user information and identify the target user according to the user information;

[0109] The voice conversion module 402 is used to convert the user agreement text of the target user into voice to obtain the agreement voice;

[0110] An interactive speech recognition module 403 is used to collect the user interactive speech corresponding to the protocol speech and recognize the interactive speech text of the user interactive speech;

[0111] A protocol explanation content generating module 404 is used to calculate the protocol explanation content of the user interactive speech according to the interactive speech text;

[0112] The auxiliary reading data collection module 405 is used to collect interactive data according to the protocol explanation content to obtain the auxiliary reading data of the target user.

[0113] In one embodiment, the target user identification module 401, when identifying the target user according to the user information, is used to:

[0114] Classifying the user information to obtain information to be identified for identifying user categories;

[0115] Perform information identification based on the information to be identified to obtain user category information;

[0116] The target user is determined according to the user category information.

[0117] In one embodiment, the voice conversion module 402, when performing voice conversion on the user agreement text of the target user to obtain the agreement voice, is used to:

[0118] The user agreement text is divided into text characters to obtain text characters:

[0119] Performing character feature encoding on the text characters to obtain a character code corresponding to each of the text characters;

[0120] Calculate the Mel spectrum corresponding to each of the text characters according to the character encoding;

[0121] The Mel spectrum is converted into speech to obtain the protocol speech corresponding to the user protocol text.

[0122] In one embodiment, the speech conversion module 402, when calculating the Mel spectrum corresponding to each character in the text according to the character encoding, is used to:

[0123] Performing short-time Fourier transform on the character code to obtain power spectrum density;

[0124] Calculate the Mel filter spectrum of the character code according to the power spectrum density;

[0125] Logarithmically compress the Mel filter spectrum to obtain Mel spectrum corresponding to the characters in the text.

[0126] In one embodiment, the interactive speech recognition module 403, when performing the recognition of the interactive speech text of the user interactive speech, is used to:

[0127] Performing speech pre-emphasis and speech noise reduction processing on the user interaction speech to obtain an enhanced speech signal;

[0128] Performing signal division on the enhanced speech signal to obtain a speech signal sequence;

[0129] Extracting speech features of the user interaction speech according to the speech signal sequence;

[0130] The voice feature is signal decoded to obtain an interactive voice text of the user interactive voice.

[0131] In one embodiment, the protocol explanation content generating module 404, when calculating the protocol explanation content of the user interactive speech according to the interactive speech text, is used to:

[0132] Determine the real-time protocol text corresponding to the protocol voice, compare the protocol text with the interactive voice text for consistency, and obtain a comparison result;

[0133] When the comparison result is consistent, searching the preset knowledge base for the protocol explanation content of the user interaction voice;

[0134] When the comparison result is inconsistent, the protocol explanation content of the user interaction voice is retrieved from the user protocol text.

[0135] In one embodiment, the auxiliary reading data collection module 405, when collecting interactive data according to the protocol to explain the content and obtain the auxiliary reading data of the target user, is used to:

[0136] Generate explanation voice according to the explanation content of the protocol;

[0137] Invoking a preset user interaction component according to the explanation voice, and collecting component interaction data of the user interaction component;

[0138] The component interaction data, the interaction voice and the protocol explanation content are collected to obtain auxiliary reading data of the target user.

[0139] The present invention provides a document auxiliary reading device, which can identify the target user and specifically identify the target user who needs auxiliary reading, help them to better understand, and improve the effect of protocol document reading; perform voice conversion on the user protocol text of the target user, and perform voice reading of the user protocol text in combination with the protocol voice, so as to improve the effect of protocol text auxiliary reading; collect user interactive voice, obtain the possible problems of the target user in real time, and the protocol explanation content of the user interactive voice can solve the problems raised by the target user in a targeted manner, so as to effectively improve the effect of auxiliary reading; collect interactive data according to the protocol explanation content, obtain the auxiliary reading data of the target user, and provide evidence in the future to avoid business disputes and business conflicts, so as to further improve the effect of protocol document auxiliary reading.

[0140] The specific definition of the document assisted reading device can be found in the definition of the document assisted reading method above, which will not be repeated here. Each module in the above document assisted reading device can be implemented in whole or in part by software, hardware and a combination thereof. Each of the above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0141] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 5 As shown. The computer device includes a processor, a memory, a network interface and a database connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile and / or volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external client through a network connection. When the computer program is executed by the processor, it realizes the functions or steps of a file auxiliary reading method server side.

[0142] In one embodiment, a computer device is provided. The computer device may be a client, and its internal structure diagram may be as follows: Figure 6 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external server through a network connection. When the computer program is executed by the processor, it realizes the functions or steps of the client side of a file auxiliary reading method

[0143] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:

[0144] Acquire user information, and identify a target user according to the user information;

[0145] Performing voice conversion on the user agreement text of the target user to obtain the agreement voice;

[0146] Collecting the user interaction voice corresponding to the protocol voice, and identifying the interactive voice text of the user interaction voice;

[0147] Calculate the protocol explanation content of the user interactive speech according to the interactive speech text;

[0148] The interactive data is collected according to the protocol to explain the content and obtain auxiliary reading data of the target user.

[0149] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0150] Acquire user information, and identify a target user according to the user information;

[0151] Performing voice conversion on the user agreement text of the target user to obtain the agreement voice;

[0152] Collecting the user interaction voice corresponding to the protocol voice, and identifying the interactive voice text of the user interaction voice;

[0153] Calculate the protocol explanation content of the user interactive speech according to the interactive speech text;

[0154] The interactive data is collected according to the protocol to explain the content and obtain auxiliary reading data of the target user.

[0155] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can refer to the relevant descriptions on the server side and the client side in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0156] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0157] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0158] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

[0159] It should be noted that if software tools or components other than those of our company appear in the embodiments of the present application, they are only used for illustrative purposes and do not represent actual use.

Claims

1. A file auxiliary reading method, characterized in that: include: Acquire user information, and identify a target user according to the user information; Performing voice conversion on the user agreement text of the target user to obtain the agreement voice; Collecting the user interaction voice corresponding to the protocol voice, and identifying the interactive voice text of the user interaction voice; Calculate the protocol explanation content of the user interactive speech according to the interactive speech text; The interactive data is collected according to the protocol to explain the content and obtain auxiliary reading data of the target user.

2. The document assisted reading method according to claim 1, characterized in that: The identifying the target user according to the user information includes: Classifying the user information to obtain information to be identified for identifying user categories; Perform information identification based on the information to be identified to obtain user category information; The target user is determined according to the user category information.

3. The document auxiliary reading method according to claim 1, characterized in that: The step of converting the user agreement text of the target user into speech to obtain the agreement speech includes: The user agreement text is divided into text characters to obtain text characters: Performing character feature encoding on the text characters to obtain a character code corresponding to each of the text characters; Calculate the Mel spectrum corresponding to each of the text characters according to the character encoding; The Mel spectrum is converted into speech to obtain the protocol speech corresponding to the user protocol text.

4. The document auxiliary reading method according to claim 3, characterized in that: The calculating, according to the character encoding, the Mel spectrum corresponding to each character in the text comprises: Performing short-time Fourier transform on the character code to obtain power spectrum density; Calculate the Mel filter spectrum of the character code according to the power spectrum density; Logarithmically compress the Mel filter spectrum to obtain Mel spectrum corresponding to the characters in the text.

5. The document auxiliary reading method according to claim 1, characterized in that: The identifying the interactive voice text of the user interactive voice includes: Performing speech pre-emphasis and speech noise reduction processing on the user interaction speech to obtain an enhanced speech signal; Performing signal division on the enhanced speech signal to obtain a speech signal sequence; Extracting speech features of the user interaction speech according to the speech signal sequence; The voice feature is signal decoded to obtain an interactive voice text of the user interactive voice.

6. The document auxiliary reading method according to claim 1, characterized in that: The calculating the protocol explanation content of the user interactive speech according to the interactive speech text includes: Determine the real-time protocol text corresponding to the protocol voice, compare the protocol text with the interactive voice text for consistency, and obtain a comparison result; When the comparison result is consistent, searching the preset knowledge base for the protocol explanation content of the user interaction voice; When the comparison result is inconsistent, the protocol explanation content of the user interaction voice is retrieved from the user protocol text.

7. The document auxiliary reading method according to claim 1, characterized in that: The step of collecting interactive data according to the protocol to explain the content and obtain auxiliary reading data of the target user includes: Generate explanation voice according to the explanation content of the protocol; Invoking a preset user interaction component according to the explanation voice, and collecting component interaction data of the user interaction component; The component interaction data, the interaction voice and the protocol explanation content are collected to obtain auxiliary reading data of the target user.

8. A document auxiliary reading device, characterized in that: include: A target user identification module, used to obtain user information and identify the target user according to the user information; A voice conversion module, used for performing voice conversion on the user agreement text of the target user to obtain the agreement voice; An interactive speech recognition module, used to collect the user interactive speech corresponding to the protocol speech, and recognize the interactive speech text of the user interactive speech; A protocol explanation content generation module, used for calculating the protocol explanation content of the user interactive speech according to the interactive speech text; The auxiliary reading data collection module is used to collect interactive data according to the protocol explanation content to obtain the auxiliary reading data of the target user.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the document assisted reading method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the document assisted reading method according to any one of claims 1 to 7 is implemented.

Images