Information screening method and device, wearable equipment and storage medium

Abstract

The invention provides an information screening method and device, wearable equipment and a storage medium, and belongs to the technical field of data processing.The method comprises the steps that a first sensitive word library corresponding to a target text is determined; the first sensitive word bank is a sensitive word bank for a specified field; determining a target information screening strategy based on an application scene of a target text, the first sensitive word bank and the second sensitive word bank, and performing information screening on the target text based on the target information screening strategy to obtain an information screening result; the second sensitive word bank is a sensitive word bank for multiple fields; and if the information screening result shows that the target text does not contain the sensitive words, converting the target text into an audio signal, and playing the audio signal. According to the information screening method and device, the wearable equipment and the storage medium provided by the invention, the sensitive word screening accuracy and screening efficiency can be improved.

Description

Technical Field

[0001] The present disclosure belongs to the field of data processing technology, and more specifically, relates to an information screening method and device, a wearable device, and a storage medium. Background Art

[0002] With the development of information technology, users can get news from various content publishing platforms (such as podcast platforms, social media platforms, etc.), so they can access a large amount of information in different fields and on different topics. Among them, some words that violate the law, social ethics and values ​​may make users feel uncomfortable, disgusted or even hurt. Before the content is released, by screening sensitive words, users can be provided with high-quality content that better meets their expectations, thereby improving user satisfaction and loyalty.

[0003] The accuracy and efficiency of existing sensitive word screening methods need to be improved. Summary of the invention

[0004] The purpose of the present disclosure is to provide an information screening method and apparatus, a wearable device, and a storage medium to improve the accuracy and efficiency of sensitive word screening.

[0005] A first aspect of the embodiments of the present disclosure provides an information screening method, comprising: Determine a first sensitive word library corresponding to the target text; the first sensitive word library is a sensitive word library for a specified field; Determine a target information screening strategy based on the application scenario of the target text, the first sensitive word library and the second sensitive word library, and screen the target text based on the target information screening strategy to obtain an information screening result; the second sensitive word library is a sensitive word library for multiple fields; If the information screening result shows that the target text does not contain sensitive words, the target text is converted into an audio signal and the audio signal is played.

[0006] A second aspect of the embodiments of the present disclosure provides an information screening device, including: A first acquisition module is used to determine a first sensitive word library corresponding to the target text; the first sensitive word library is a sensitive word library for a specified field; An information screening module, configured to determine a target information screening strategy based on an application scenario of a target text, the first sensitive word library, and the second sensitive word library, and to screen the target text based on the target information screening strategy to obtain an information screening result; the second sensitive word library is a sensitive word library for multiple fields; The text output module converts the target text into an audio signal and plays the audio signal if the information screening result shows that the target text does not contain sensitive words.

[0007] According to a third aspect of an embodiment of the present disclosure, a wearable device is provided, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of the above-mentioned information screening method when executing the computer program.

[0008] According to a fourth aspect of the embodiments of the present disclosure, 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 information screening method are implemented.

[0009] The information screening method and device, wearable device, and storage medium provided by the embodiments of the present disclosure have the following beneficial effects: In the disclosed embodiment, the target text is a text that needs to be screened for sensitive words. The first sensitive word library can be customized based on a specified field (that is, the field to which the target text belongs). By customizing the target text through the first sensitive word library, sensitive content related to the specified field can be accurately identified and filtered. The second sensitive word library can adopt the sensitive word library provided by a third-party platform. The third-party platform usually has a rich sensitive word library, which is obtained through a large amount of data training and manual sorting. It can quickly perform preliminary screening on the content of the target text, saving a lot of time and effort.

[0010] Therefore, the embodiment of the present disclosure combines the first sensitive word library of the specified field with the second general sensitive word library to screen sensitive words, which can minimize the omission of sensitive words and improve the screening efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0012] Figure 1 A flowchart of an information screening method provided by an embodiment of the present disclosure; Figure 2 A schematic diagram of a target topic prompt word selection process provided by an embodiment of the present disclosure; Figure 3 A schematic diagram of a target text generation process provided by an embodiment of the present disclosure; Figure 4 A structural block diagram of an information screening device provided by an embodiment of the present disclosure; Figure 5A schematic block diagram of a wearable device provided in one embodiment of the present disclosure. DETAILED DESCRIPTION

[0013] In the following description, specific details such as specific system structures and technologies are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present disclosure. However, it should be clear to those skilled in the art that the present disclosure may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid obstructing the description of the present disclosure with unnecessary details.

[0014] In order to make the purpose, technical solutions and advantages of the present disclosure more clear, specific embodiments will be described below in conjunction with the accompanying drawings.

[0015] Please refer to Figure 1 , Figure 1 A flowchart of an information screening method provided by an embodiment of the present disclosure, the method comprising: S101: Determine a first sensitive word library corresponding to a target text; the first sensitive word library is a sensitive word library for a specified field.

[0016] In this embodiment, the target text is a text that needs to be screened for sensitive words, and the field to which the target text belongs can be determined as the designated field.

[0017] Different designated fields have their own unique categories of sensitive words. For example, in the medical field, there may be sensitive expressions such as unauthorized drug promotion and illegal medical practice; in the financial field, there may be related sensitive words such as illegal fundraising and financial fraud. By building the first sensitive word library for the designated field, we can more accurately capture potential sensitive words related to the content in the field, making the screening more in line with the actual situation of the target text. Specifically, for a designated field, we first collect a large amount of text data in that field. Taking the news media field as an example, we can collect a large number of news reports, commentary articles and other text content. In the field of science and technology, we can collect academic papers, technical blogs, product manuals and other texts. The above text data will serve as the basis for mining sensitive words.

[0018] Then, the above text data can be analyzed using natural language processing technology. For example, topic models (such as LDA - Latent Dirichlet Allocation) can be used to mine topic-related words in the text. For example, in the text in the field of information technology, the topic model can be used to mine words related to topics such as "network security vulnerability exploitation", "illegal data theft", and "malware transmission channels" as candidate words for building the first sensitive word library.

[0019] Considering that the automatically mined words may contain misjudgments or inaccuracies, manual review is required. Manual review can screen and adjust the automatically mined words based on domain knowledge and actual application scenarios, and finally determine the words in the first sensitive word library. For example, when reviewing the words mined in the field of information technology, words such as "hacker attack tool tutorial" and "new methods of online fraud revealed" that are related to crimes and endangering network security are retained.

[0020] S102: Determine a target information screening strategy based on the application scenario of the target text, the first sensitive word library, and the second sensitive word library, and screen the target text based on the target information screening strategy to obtain an information screening result; the second sensitive word library is a sensitive word library for multiple fields.

[0021] In this embodiment, the second sensitive word library can be a sensitive word library provided by a third-party platform. The third-party platform usually has a rich sensitive word library, which covers words that are generally considered sensitive in many fields, such as politically sensitive words, vulgar words, terrorist words, etc. At the same time, the sensitive word library provided by the third-party platform is obtained through a large amount of data training and manual sorting, and can quickly perform preliminary screening of the content of the target text, saving a lot of time and energy.

[0022] Considering the professionalism and particularity of the specific field in which the target text is located, the first sensitive word library can be used for targeted screening; at the same time, the target text must also comply with general social ethics and legal and regulatory requirements, and the second sensitive word library can be used for universal inspection. Two rounds of information screening of the target text based on the first sensitive word library and the second sensitive word library can improve the accuracy and comprehensiveness of the target text screening.

[0023] S103: If the information screening result shows that the target text does not contain sensitive words, the target text is converted into an audio signal and the audio signal is played.

[0024] In this embodiment, if the information screening results after the above two rounds of information screening show that the target text does not contain sensitive words, the target text can be output. The target text can be published on the website in text form, or it can be converted into a script suitable for podcast recording and published in podcast form.

[0025] For example, for an article in the field of information technology, the corresponding first sensitive word library may be words such as "illegal network intrusion", "unauthorized data acquisition", "intentional introduction of code injection vulnerabilities", etc., and the second sensitive word library includes common sensitive words such as "sensitive word W1", "sensitive word W2", "sensitive word W3", etc. If the article does not contain sensitive words in the first and second sensitive word libraries, then the information screening results show that the target text has passed the review, and this information technology article can be output for readers to read.

[0026] Through the above two rounds of information screening, we can not only ensure that the content of the article is compliant in the field of information technology, but also prevent information technology articles from being maliciously inserted or externally contaminated during the editing process, thereby avoiding adverse effects on society.

[0027] From the above, it can be concluded that this embodiment uses the first sensitive word library to perform customized screening of the target text, which can accurately identify and filter sensitive content related to the specified field; the second sensitive word library can adopt the sensitive word library provided by a third-party platform. The third-party platform usually has a rich sensitive word library, which is obtained through a large amount of data training and manual sorting. It can quickly perform preliminary screening of the content of the target text, saving a lot of time and energy.

[0028] Therefore, this embodiment combines the first sensitive word library of the specified field with the second general sensitive word library to screen sensitive words, which can minimize the omission of sensitive words and improve the screening efficiency.

[0029] In one embodiment of the present disclosure, a target information screening strategy is determined based on an application scenario of a target text, a first sensitive word library, and a second sensitive word library, including: If the application scenario of the target text belongs to the first type of application scenario, the first information screening strategy is determined as the target information screening strategy; the first type of application scenario includes a government document publishing platform, a financial regulatory document and reporting platform; The target text is screened based on the target information screening strategy to obtain information screening results, including: Filtering information of the target text based on the second sensitive word library to obtain a first evaluation value; If the first evaluation value is greater than the first threshold, the information screening result is determined as containing sensitive words; If the first evaluation value is less than or equal to the first threshold, information screening is performed on the target text based on the first sensitive word library to obtain a second evaluation value; If the second evaluation value and the first evaluation value are both less than or equal to the first threshold, the information screening result is determined to contain no sensitive words.

[0030] In this embodiment, a specific implementation method for information screening of the target text based on the first sensitive word library and the second sensitive word library is provided. This implementation method can be applied to scenarios with strict requirements on content supervision, such as government document publishing platforms, financial regulatory documents and reporting platforms, etc.

[0031] The first evaluation value is used to represent the probability that the target text contains a common sensitive word. The larger the first evaluation value, the greater the probability that the target text contains a common sensitive word. When the probability is greater than the set value, that is, the first evaluation value is greater than the first threshold, it can be determined that the target text contains a common sensitive word. The first threshold is a preset constant, and those skilled in the art can determine the specific value of the first threshold according to actual needs.

[0032] Optionally, the target text can be compared with the second sensitive word library to determine which common sensitive words are included in the target text, and the probability of the common sensitive words appearing (that is, the proportion of the number of common sensitive words to the total number of words in the target text) can be counted, and the size of the first evaluation value can be determined based on the probability of the common sensitive words appearing. The greater the probability of the common sensitive words appearing, the higher the first evaluation value. Specifically, the probability range of the common sensitive words appearing can be divided into several different probability range intervals, and a corresponding score can be set for each interval, and the corresponding score can be used as the first evaluation value.

[0033] The second evaluation value is used to represent the probability that the target text contains sensitive words in the specified field. The larger the second evaluation value, the greater the probability that the target text contains sensitive words in the specified field. When the probability is greater than the set value, that is, the second evaluation value is greater than the first threshold, it can be determined that the target text contains sensitive words in the specified field.

[0034] Similarly, the target text can be compared with the first sensitive word library to determine which sensitive words in the specified field are contained in the target text, and the probability of the occurrence of sensitive words in the specified field of sensitive words (that is, the proportion of the number of sensitive words in the specified field to the total number of words in the target text) can be counted. The size of the second evaluation value is determined based on the probability of the occurrence of sensitive words in the specified field. The greater the probability of the occurrence of sensitive words in the specified field, the higher the second evaluation value. Specifically, the probability range of the occurrence of sensitive words in the specified field can be divided into several different probability range intervals, and a corresponding score is set for each interval, and the corresponding score is used as the second evaluation value. .

[0035] Taking the financial regulatory documents and reporting platform as an example, the financial industry is strictly regulated. For documents issued by financial regulatory agencies, reports submitted by financial institutions, and other contents, sensitive words such as "illegal fundraising" and "financial fraud" cannot appear. This embodiment first uses the first sensitive word library to perform preliminary information screening to obtain a first evaluation value. If the first evaluation value is greater than the first threshold, it indicates that the target text contains common sensitive words, and it is directly judged as containing sensitive words without subsequent judgment. Only when the first evaluation value is less than or equal to the first threshold, information screening is performed for the financial field, and the target text is further analyzed in depth to dig out potential sensitive words that have not been obviously discovered in the preliminary screening, effectively prevent the spread of false information or illegal content, and maintain the order of the financial market.

[0036] From the above, it can be concluded that the hierarchical screening mechanism of this embodiment, which first performs preliminary screening based on the second sensitive word library to obtain the first evaluation value, and then determines whether to perform re-screening to obtain the second evaluation value based on the comparison between the first evaluation value and the first threshold, can effectively improve the screening efficiency. In scenarios with strict requirements for content supervision, preliminary screening can quickly exclude target texts that obviously contain common sensitive words, reduce the number of texts that need to be re-screened later, and save time and computing resources.

[0037] In one embodiment of the present disclosure, a target information screening strategy is determined based on an application scenario of a target text, a first sensitive word library, and a second sensitive word library, including: If the application scenario of the target text belongs to the second type of application scenario, the second information screening strategy is determined as the target information screening strategy; the second type of application scenario includes comprehensive social media platforms and news information aggregation platforms; The target text is screened based on the target information screening strategy to obtain information screening results, including: Filtering information of the target text based on the second sensitive word library to obtain a first evaluation value; Filtering information of the target text based on the first sensitive word library to obtain a second evaluation value; Performing a weighted summation on the first evaluation value and the second evaluation value to obtain a third evaluation value; If the third evaluation value is less than or equal to the second threshold, the information screening result is determined to not contain sensitive words.

[0038] In this embodiment, another specific implementation method of screening information of the target text based on the first sensitive word library and the second sensitive word library is provided. This implementation method can be applied to general content platforms for the general public, such as comprehensive social media platforms, news information aggregation platforms, etc.

[0039] The calculation method of the first evaluation value and the second evaluation value is the same as that of the above embodiment, which will not be repeated here. Based on the first evaluation value and the second evaluation value, this embodiment performs weighted summation on the first evaluation value and the second evaluation value, and determines the information screening result based on the result of the weighted summation.

[0040] Taking the comprehensive social media platform as an example, considering the rich and diverse content themes on the comprehensive social media platform, the tolerance of different themes to sensitive words may be different. In some emerging content fields or innovative content expressions, general sensitive words or designated field sensitive words may be involved to a certain extent due to the breakthrough of traditional expressions. For example, the introduction copy of some works of art may use some bold and avant-garde expressions, which may contain some seemingly sensitive words, but from the perspective of artistic innovation and expression, they have their unique value. At this time, the judgment criteria for general sensitive words or designated field sensitive words can be appropriately lowered. As long as the third evaluation value after the weighted sum of the first evaluation value and the second evaluation value is within the acceptable range (that is, less than or equal to the second threshold), it can be determined that the target text does not contain sensitive words, so that the platform can better adapt to the characteristics of various different themes and avoid stifling valuable content creation due to overly strict restrictions on a certain type of sensitive words.

[0041] From the above, it can be concluded that this embodiment determines the information screening result based on the weighted sum of the first evaluation value and the second evaluation value, which can comprehensively consider various characteristics of the target text content, more flexibly judge whether the target text content contains sensitive words, and adapt to the social media environment with rich and diverse content topics.

[0042] In one embodiment of the present disclosure, the information screening method further includes: The weight of the first evaluation value and the weight of the second evaluation value are determined based on the attribute information of the target text; the attribute information of the target text includes the domain information and the audience information of the target text.

[0043] In this embodiment, taking into account that different fields (or topics) and audiences have different focuses on sensitive words, the weights of the first evaluation value and the second evaluation value are determined based on the field information and audience information of the target text, which can more accurately distinguish which sensitive words really need attention and reduce the risks of misjudgment and missed judgment.

[0044] Specifically, the weight of the first evaluation value and the weight of the second evaluation value determined based on the attribute information of the target text can be described in detail as follows: In response to the domain information of the target text belonging to the first category of domain information, the weight of the first evaluation value is set to a first value, and the weight of the second evaluation value is set to a second value; the first value is less than the second value, and the first category of domain information is domain information corresponding to a subject in a professional and technical field; In response to the target text's domain information belonging to the second category of domain information, the weight of the first evaluation value is set to the second value, and the weight of the second evaluation value is set to the first value; the second category of domain information is domain information corresponding to the social life domain; The weights of the first evaluation value and the second evaluation value are adjusted based on the audience information.

[0045] In this embodiment, the topics corresponding to the first category of domain information may include professional and technical fields such as natural sciences, social sciences, and engineering technology. In texts in the professional and technical field, it is necessary to highlight the weight of professional and technical related sensitive words in order to accurately capture sensitive content such as technology infringement and false scientific research results publicity that may seriously affect the development of the professional field. Therefore, a higher weight (for example, 70%) can be given to the second evaluation value in order to more effectively capture and prevent sensitive words in the professional field.

[0046] The topics corresponding to the second category of domain information may include social life areas such as life services, culture and entertainment, and news and information. The social life area covers all aspects of people's daily life, such as social interaction, family, consumption, leisure and entertainment, etc. The content of this type of topic has a wide audience, a relatively popular style, and involves a lot of considerations in terms of social morality, public order and good customs, so a higher weight (for example, 70%) is given to the first evaluation value, and those common sensitive words that may cause greater social impact can be excluded first.

[0047] On this basis, the audience information can be divided into three age groups: minors, adults, and the elderly. Different adjustment parameters are assigned to the audience information of each age group, and the weights of the first evaluation value and the second evaluation value are adjusted based on the adjustment parameters corresponding to each age group. For example, for the minor stage, in order to ensure that the content of the target text conforms to the psychological and cognitive characteristics of children, it is necessary to pay more attention to the weight of the second evaluation value. Therefore, the weight of the second evaluation value can be increased based on the first adjustment parameter (greater than 1) to prevent the appearance of adult, horror and other content. For the elderly stage, the elderly may have relatively weak ability to discern new information. Health and wellness podcasts for the elderly need to strictly screen content. Therefore, the second adjustment parameter (greater than 1) can be set to increase the weight of the second evaluation value to avoid false health product propaganda, unscientifically verified treatment methods and other content.

[0048] The sum of the weight of the first evaluation value and the weight of the second evaluation value is equal to 1. Therefore, after the weight of the second evaluation value is adjusted based on the audience information, the size of the first evaluation value needs to be re-determined based on the adjusted second evaluation value.

[0049] From the above, it can be concluded that for the target text in the professional and technical field, this embodiment can effectively screen out sensitive words that affect the accuracy and standardization of professional and technical knowledge by highlighting the weight of the second evaluation value, thereby ensuring the quality of professional content. In the field of social life, giving a higher weight to the first evaluation value helps to exclude vulgar, politically sensitive and other content that violates social order and good customs, and create a healthy and civilized content environment for the general audience. On this basis, taking into account the needs of special audiences such as minors and the elderly, adjusting the weight of the second evaluation value based on audience information can avoid various risks caused by special audience groups being exposed to inappropriate content, further improving the effectiveness of sensitive word screening.

[0050] refer to Figure 2 In one embodiment of the present disclosure, the information screening method further includes: In response to receiving a play instruction from the first user for the first time, arbitrarily selecting a target theme prompt word from a plurality of first theme prompt words input by the first user, and generating a target text based on the target theme prompt word and the preference information of the first user; In response to receiving a playback instruction from the first user for another time, selecting a target theme prompt word based on weights of a plurality of first theme prompt words; the weights of the first theme prompt words are obtained based on historical playback data; A target text is generated based on the target topic prompt word and the preference information of the first user.

[0051] The method of this embodiment can be applied to wearable devices such as smart audio glasses. By converting the target text into an audio signal output, such as outputting it in the form of a podcast, users can obtain information more conveniently. For scenarios where it is inconvenient to read text, such as when driving, exercising, or doing other things, users can understand related content by listening to audio.

[0052] Furthermore, users can pre-set topic prompts of their interest, and can also further input preference information to refine requirements for the style and depth of the generated text. In this way, when users want to listen to podcasts, they do not need to input specific requirements for the playback content. The wearable device can automatically play the audio signal according to the topic prompts and preference information input by the user. For example, a user inputs the topic prompt word "technology" and the preference information is a concise style. Based on this, the wearable device can generate a concise style target text about the latest progress in science and technology, so that users feel that the content is tailored for them, thereby improving users' satisfaction and acceptance of the content.

[0053] Specifically, taking the first user as an example, when the wearable device receives the first user's play instruction for the first time, the wearable device will randomly select a target topic prompt word from the multiple first topic prompt words input by the first user, and then generate the target text based on the target topic prompt word and the first user's preference information. Since the topic prompt words input by the user represent the topic they are currently interested in, and the preference information further refines the requirements for the style and depth of the generated text, the generated target text can meet the user's expectations and interests to a certain extent.

[0054] If the first user has issued a play instruction, the wearable device will analyze the popularity or importance of each topic prompt word based on the first user's past play behavior, and then more specifically select the topic prompt word to generate the target text this time. For example, if a topic prompt word appears frequently in the user's previous play history or the user has listened to the podcast on this topic for a long time, the weight of the topic prompt word will be relatively high, and it is more likely to be selected as the topic prompt word for generating the target text this time.

[0055] It can be concluded from the above that this embodiment fully considers the subject prompt words and preference information of the first user when generating the target text, so that the generated content is highly consistent with the user's personal interests, reading style and other needs, thereby improving the user's satisfaction and acceptance of the played content.

[0056] In one embodiment of the present disclosure, the information screening method further includes: The weight of the first topic prompt word is adjusted based on the playback duration of the target audio signal and the feedback information of the first user on the target audio signal; the target audio signal is an audio signal corresponding to the target text.

[0057] In this embodiment, the target audio signal is obtained based on audio conversion of the target text, and the feedback information of the first user on the target audio signal may include the first user's like / collection operation information on the target audio signal and the interaction rounds on the target audio signal. When the first user performs a like / collection operation on the target audio signal, it indicates that the user has a positive emotional response to the audio content during the listening process and wants to listen to it again in the future; the interaction round refers to the number of times the user interacts with the audio content during the process of listening to the target audio signal, such as the number of questions asked about the audio content, and the number of interaction rounds reflects the user's attention to the audio content.

[0058] This embodiment takes into account that the playback time of the target audio signal can reflect the user's attention and interest in the target audio signal. If the playback time of the target audio signal is long, it indicates that the user is more interested in the corresponding topic and is willing to spend more time listening to it, so the weight of the corresponding first theme word will increase accordingly. Conversely, if the playback time is very short, it indicates that the user is less interested in the topic, and the weight of the corresponding first theme word will decrease.

[0059] Specifically, you can pre-set the baseline value of the playback duration. and the third proportionality factor On this basis, based on the playback time of the target audio signal and benchmark value The difference between the values ​​of the first topic prompt word is used to calculate the change in the weight of the first topic prompt word. Specifically, the following formula can be used to calculate the change in the weight of the first topic prompt word: : .

[0060] At the same time, the change amount of the weight of the first topic prompt word is calculated based on the feedback information of the first user for the target audio signal :

[0061] in, Indicates the number of like / collection operations. Indicates no like / collection operation. Indicates only like or collection operations. Indicates that there is both a like operation and a favorite operation. represents the number of interaction rounds, that is, the number of times the first user asks a question, is the preset fourth proportional coefficient, is the preset fifth proportional coefficient.

[0062] In getting the change and the amount of change Based on and the amount of change The adjusted weight of the first topic prompt word is obtained by adding it to the weight of the first topic prompt word.

[0063] From the above, it can be concluded that this embodiment adjusts the weight of the first topic prompt word by comprehensively considering the playback time and feedback information, so as to more accurately grasp the first user's preference for different topics. The subsequent selection of the target topic prompt word based on the weight can better meet the user's interests and allow the user to have a better experience during use.

[0064] refer to Figure 3In one embodiment of the present disclosure, the information screening method further includes: Calculate the semantic similarity between the target text and the target topic cue words; Calling a corresponding quantitative description template based on user preference information; Perform statistical analysis on the target text, fill in the quantitative description template based on the results of the statistical analysis, and obtain a quantitative description of the user's preference information; determining a first similarity between the target text and the user preference information based on the quantitative description of the user preference information; If the semantic similarity and the first similarity are both greater than the similarity threshold, the target text is output.

[0065] In this embodiment, the target text and the target topic prompt word can be converted into vector representations respectively, and then the distance between the vectors (such as cosine distance) is calculated to determine the semantic similarity between the target text and the target topic prompt word. If the semantic similarity is greater than or equal to the similarity threshold, it indicates that the target text is close to the user's expected topic prompt word in terms of subject content, otherwise, if the semantic similarity is less than the similarity threshold, it indicates that the target text is not close to the user's expected topic prompt word in terms of subject content.

[0066] According to the user's preference information, a corresponding quantitative description template can be preset. For example, if the user's preference information on style is easy to understand, the corresponding quantitative description template can be: the average vocabulary difficulty value is [ ]; the average sentence length is: [ ].

[0067] On this basis, statistical analysis of the target text can be performed to pre-build a vocabulary difficulty dictionary and a difficulty coefficient threshold, and assign a corresponding difficulty coefficient to each vocabulary (for example, the coefficient of common simple vocabulary is 1-3, the coefficient of medium difficulty vocabulary is 4-6, and the coefficient of rare or professional vocabulary is 7-10). The sum of the difficulty coefficients of all vocabulary in the target text is counted and then divided by the total number of vocabulary in the target text to obtain the average vocabulary difficulty value. The first difference between the difficulty coefficient threshold and the average vocabulary difficulty value is calculated, and then the first difference is multiplied by the preset sixth proportional coefficient to obtain the second similarity.

[0068] At the same time, a sentence length threshold can be preset, and the sum of the number of words in all sentences in the target text is calculated and divided by the total number of sentences to obtain the average sentence length. The second difference between the sentence length threshold and the average sentence length is calculated, and then the second difference is multiplied by the preset seventh proportional coefficient to obtain the third similarity.

[0069] On the basis of obtaining the second similarity and the third similarity, the first similarity can be obtained by weighted summing the second similarity and the third similarity.

[0070] If both the semantic similarity and the first similarity are greater than the similarity threshold, it indicates that the target text meets the user's expectations. At this time, the target text can be output, for example, the target text is stored in the proxy content pool and waits for playback. Otherwise, the target text is discarded.

[0071] From the above, it can be concluded that this embodiment can ensure that the generated target text is highly consistent with the user's interests in terms of subject matter by calculating the semantic similarity. At the same time, based on the quantitative description of the user's preference information and the calculation of the first similarity, the user's preferences in content style, depth, etc. can be considered, so that the output target text is not only appropriate in subject matter, but also can meet the user's personalized needs in all details, further improving the overall matching degree between content and user needs.

[0072] Corresponding to the information screening method of the above embodiment, Figure 4 This is a structural block diagram of an information screening device provided by an embodiment of the present disclosure. For ease of explanation, only the parts related to the embodiment of the present disclosure are shown. Figure 4 The information screening device 20 includes: a first acquisition module 21, an information screening module 22 and a text output module 23. The first acquisition module 21 is used to determine a first sensitive word library corresponding to the target text; the first sensitive word library is a sensitive word library for a specified field; The information screening module 22 is used to determine a target information screening strategy based on the application scenario of the target text, the first sensitive word library and the second sensitive word library, and screen the target text based on the target information screening strategy to obtain an information screening result; the second sensitive word library is a sensitive word library for multiple fields; The text output module 23 converts the target text into an audio signal and plays the audio signal if the information screening result shows that the target text does not contain sensitive words.

[0073] In one embodiment of the present disclosure, the information screening module 22 is specifically used to: If the application scenario of the target text belongs to the first type of application scenario, the first information screening strategy is determined as the target information screening strategy; the first type of application scenario includes a government document publishing platform, a financial regulatory document and reporting platform; The target text is screened based on the target information screening strategy to obtain information screening results, including: Filtering information of the target text based on the second sensitive word library to obtain a first evaluation value; If the first evaluation value is greater than the first threshold, the information screening result is determined as containing sensitive words; If the first evaluation value is less than or equal to the first threshold, information screening is performed on the target text based on the first sensitive word library to obtain a second evaluation value; If the second evaluation value and the first evaluation value are both less than or equal to the first threshold, the information screening result is determined to contain no sensitive words.

[0074] In one embodiment of the present disclosure, the information screening module 22 is specifically used to: If the application scenario of the target text belongs to the second type of application scenario, the second information screening strategy is determined as the target information screening strategy; the second type of application scenario includes comprehensive social media platforms and news information aggregation platforms; The target text is screened based on the target information screening strategy to obtain information screening results, including: Filtering information of the target text based on the second sensitive word library to obtain a first evaluation value; Filtering information of the target text based on the first sensitive word library to obtain a second evaluation value; Performing a weighted summation on the first evaluation value and the second evaluation value to obtain a third evaluation value; If the third evaluation value is less than or equal to the second threshold, the information screening result is determined to not contain sensitive words.

[0075] In one embodiment of the present disclosure, the information screening module 22 is further configured to: The weight of the first evaluation value and the weight of the second evaluation value are determined based on the attribute information of the target text; the attribute information of the target text includes the domain information and the audience information of the target text.

[0076] In one embodiment of the present disclosure, the first acquisition module 21 is specifically used to: In response to receiving a play instruction from the first user for the first time, arbitrarily selecting a target theme prompt word from a plurality of first theme prompt words input by the first user, and generating a target text based on the target theme prompt word and the preference information of the first user; In response to receiving a playback instruction from the first user for another time, selecting a target theme prompt word based on weights of a plurality of first theme prompt words; the weights of the first theme prompt words are obtained based on historical playback data; A target text is generated based on the target topic prompt word and the preference information of the first user.

[0077] In an embodiment of the present disclosure, the first acquisition module 21 is further configured to: The weight of the first topic prompt word is adjusted based on the playback duration of the target audio signal and the feedback information of the first user on the target audio signal; the target audio signal is an audio signal corresponding to the target text.

[0078] In one embodiment of the present disclosure, the text output module 23 is specifically used for: Calculate the semantic similarity between the target text and the target topic cue words; Calling a corresponding quantitative description template based on user preference information; Perform statistical analysis on the target text, fill in the quantitative description template based on the results of the statistical analysis, and obtain a quantitative description of the user's preference information; determining a first similarity between the target text and the user preference information based on the quantitative description of the user preference information; If the semantic similarity and the first similarity are both greater than the similarity threshold, the target text is output.

[0079] See also Figure 5 , Figure 5 A schematic block diagram of a wearable device provided by an embodiment of the present disclosure. Figure 5 The wearable device 300 in the embodiment shown may include: one or more processors 301, one or more input devices 302, one or more output devices 303 and one or more memories 304. The processors 301, input devices 302, output devices 303 and memories 304 communicate with each other via a communication bus 305. The memory 304 is used to store computer programs, which include program instructions. The processor 301 is used to execute the program instructions stored in the memory 304. The processor 301 is configured to call the program instructions to execute the functions of each module / unit in the above-mentioned device embodiments, such as Figure 4 The functions of modules 21 to 23 are shown.

[0080] It should be understood that in the embodiment of the present disclosure, the processor 301 may be a central processing unit (CPU), and the processor may also be 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 the processor may also be any conventional processor, etc.

[0081] The input device 302 may include a touch panel, a fingerprint collection sensor (for collecting the user's fingerprint information and fingerprint direction information), a microphone, etc., and the output device 303 may include a display (LCD, etc.), a speaker, etc.

[0082] The memory 304 may include a read-only memory and a random access memory, and provide instructions and data to the processor 301. A portion of the memory 304 may also include a non-volatile random access memory. For example, the memory 304 may also store information about the device type.

[0083] In a specific implementation, the processor 301, input device 302, and output device 303 described in the embodiments of the present disclosure can execute the implementation methods described in the first and second embodiments of the information screening method provided in the embodiments of the present disclosure, and can also execute the implementation methods of the wearable device described in the embodiments of the present disclosure, which will not be repeated here.

[0084] In another embodiment of the present disclosure, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by the processor, all or part of the processes in the above-mentioned embodiment method are implemented, and the computer program can also be completed by instructing the relevant hardware through the computer program. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by the processor, the steps of each of the above-mentioned method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0085] The computer-readable storage medium may be an internal storage unit of the wearable device of any of the foregoing embodiments, such as a hard disk or memory of the wearable device. The computer-readable storage medium may also be an external storage device of the wearable device, such as a plug-in hard disk equipped on the wearable device, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. Furthermore, the computer-readable storage medium may also include both an internal storage unit of the wearable device and an external storage device. The computer-readable storage medium is used to store computer programs and other programs and data required by the wearable device. The computer-readable storage medium may also be used to temporarily store data that has been output or is to be output.

[0086] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this disclosure.

[0087] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the wearable device and unit described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0088] In the several embodiments provided in the present application, it should be understood that the disclosed wearable devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces or units, or it can be an electrical, mechanical or other form of connection.

[0089] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of the present disclosure.

[0090] In addition, each functional unit in each embodiment of the present disclosure may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0091] The above are only specific embodiments 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 equivalent modifications or replacements within the technical scope disclosed in the present disclosure, and these modifications or replacements 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. An information screening method, characterized in that: include: Determine the first sensitive word library corresponding to the target text; The first sensitive word library is a sensitive word library for a specified field; Determining a target information screening strategy based on an application scenario of the target text, the first sensitive word library, and the second sensitive word library, and screening the target text based on the target information screening strategy to obtain an information screening result; The second sensitive word library is a sensitive word library for multiple fields; If the information screening result shows that the target text does not contain sensitive words, the target text is converted into an audio signal and the audio signal is played.

2. The information screening method according to claim 1, characterized in that: The target information screening strategy is determined based on the application scenario of the target text, the first sensitive word library, and the second sensitive word library, including: If the application scenario of the target text belongs to the first type of application scenario, the first information screening strategy is determined as the target information screening strategy; the first type of application scenario includes a government document publishing platform, a financial regulatory document and reporting platform; The step of performing information screening on the target text based on the target information screening strategy to obtain an information screening result includes: Performing information screening on the target text based on the second sensitive word library to obtain a first evaluation value; If the first evaluation value is greater than a first threshold, determining the information screening result as containing sensitive words; If the first evaluation value is less than or equal to a first threshold, filtering the target text based on the first sensitive word library to obtain a second evaluation value; If the first evaluation value and the second evaluation value are both less than or equal to a first threshold, the information screening result is determined to contain no sensitive words.

3. The information screening method according to claim 1, characterized in that: The target information screening strategy is determined based on the application scenario of the target text, the first sensitive word library, and the second sensitive word library, including: If the application scenario of the target text belongs to the second type of application scenario, the second information screening strategy is determined as the target information screening strategy; the second type of application scenario includes a comprehensive social media platform and a news information aggregation platform; The step of performing information screening on the target text based on the target information screening strategy to obtain an information screening result includes: Performing information screening on the target text based on a second sensitive word library to obtain a first evaluation value; Performing information screening on the target text based on the first sensitive word library to obtain a second evaluation value; Performing a weighted summation of the first evaluation value and the second evaluation value to obtain a third evaluation value; If the third evaluation value is less than or equal to the second threshold, the information screening result is determined to contain no sensitive words.

4. The information screening method according to claim 3, characterized in that: Also includes: Determining a weight of the first evaluation value and a weight of the second evaluation value based on attribute information of the target text; The attribute information of the target text includes the domain information and the audience information of the target text.

5. The information screening method according to claim 1, characterized in that: Also includes: In response to receiving a play instruction from a first user for the first time, arbitrarily selecting a target topic prompt word from a plurality of first topic prompt words input by the first user, and generating the target text based on the target topic prompt word and the preference information of the first user; In response to receiving a play instruction from the first user for another time, selecting a target theme prompt word based on weights of the plurality of first theme prompt words; The weight of the first topic prompt word is obtained based on historical playback data; The target text is generated based on the target topic prompt word and the preference information of the first user.

6. The information screening method according to claim 5, characterized in that: Also includes: adjusting the weight of the first topic prompt word based on the playback duration of the target audio signal and feedback information of the first user regarding the target audio signal; The target audio signal is an audio signal corresponding to the target text.

7. The information screening method according to claim 5, characterized in that: The information screening method further includes: Calculating the semantic similarity between the target text and the target topic prompt word; Calling a corresponding quantitative description template based on user preference information; Performing statistical analysis on the target text, and filling in the quantitative description template based on the result of the statistical analysis to obtain a quantitative description of the user preference information; Determining a first similarity between the target text and the user preference information based on the quantitative description of the user preference information; If the semantic similarity and the first similarity are both greater than a similarity threshold, the target text is output.

8. An information screening device, characterized in that: include: A first acquisition module, used to determine a first sensitive word library corresponding to the target text; The first sensitive word library is a sensitive word library for a specified field; An information screening module, configured to determine a target information screening strategy based on an application scenario of a target text, the first sensitive word library, and the second sensitive word library, and to screen the target text based on the target information screening strategy to obtain an information screening result; The second sensitive word library is a sensitive word library for multiple fields; The text output module converts the target text into an audio signal and plays the audio signal if the information screening result shows that the target text does not contain sensitive words.

9. A wearable device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

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

Images