Intelligent short message platform management and control method and system based on multi-level risk control

Abstract

This specification relates to the field of information technology, specifically to a method and system for managing an intelligent SMS platform based on multi-level risk control. The method includes: reading the text of an SMS to be sent and comparing it with a preset block keyword library; if a matching block keyword exists, the SMS to be sent is blocked; identifying the semantics of the SMS to be sent and comparing it with several preset prohibited scenarios to obtain a scenario matching degree; if the scenario matching degree is higher than a preset threshold, the SMS to be sent is blocked; marking the text corresponding to the identified semantics; reading unmarked text; if the number of characters in the unmarked text exceeds a preset character count threshold, obtaining characters with similar shapes and similar pronunciations; attempting semantic recognition based on the combination of the characters with similar shapes and similar pronunciations in the unmarked text; comparing it with several preset prohibited scenarios; if any second scenario matching degree is higher than a preset threshold, the SMS to be sent is blocked.

Description

Technical Field

[0001] Several embodiments of this specification relate to the field of information technology, specifically to a method and system for managing an intelligent SMS platform based on multi-level risk control. Background Technology

[0002] Traditional SMS filtering mechanisms primarily employ keyword matching technology, which identifies and blocks SMS messages containing specific keywords using a pre-defined sensitive word database. However, keyword matching filtering methods have significant limitations. They are ineffective against variant words disguised through homophones or similar-looking characters; and they also fail to identify SMS messages that do not directly use sensitive words but whose actual content involves violations, such as those containing niche vocabulary or suggestive language. Therefore, improvements and further research into SMS blocking technology are necessary. Summary of the Invention

[0003] This specification describes a method and system for managing an intelligent SMS platform based on multi-level risk control through several embodiments.

[0004] Firstly, the embodiments of this specification provide a method for managing an intelligent SMS platform based on multi-level risk control, including the following steps:

[0005] Read the text message to be sent, compare the text message with a preset block keyword library, and if a matching block keyword is found, block the text message to be sent.

[0006] Identify the semantics of the SMS message to be sent, compare the semantics with several preset prohibited scenarios, and obtain the scenario matching degree;

[0007] If any scenario has a matching degree higher than a preset threshold, the SMS message to be sent will be blocked.

[0008] Based on the identified semantics, mark the text in the SMS message to be sent that corresponds to the semantics;

[0009] Read unmarked text. If the number of characters in the unmarked text exceeds a preset character threshold, obtain the characters with similar shapes and similar pronunciations for each character in the unmarked text.

[0010] Based on the combination of similar-looking and similar-sounding characters in the unlabeled text, semantic recognition is attempted to obtain several semantic recognition results;

[0011] The second scene matching degree is obtained by comparing the semantic recognition result with a number of preset prohibited scenes;

[0012] If any second scenario has a matching degree higher than a preset threshold, the SMS message to be sent will be intercepted.

[0013] Secondly, embodiments of this specification provide an intelligent SMS platform management system based on multi-level risk control, including:

[0014] The first interception module reads the text message to be sent and compares it with a preset interception keyword library. If a matching interception keyword is found, the text message to be sent is intercepted.

[0015] The first identification module identifies the semantics of the SMS message to be sent, compares the semantics with a number of preset prohibited scenarios, and obtains the scenario matching degree.

[0016] The second interception module intercepts the SMS message to be sent if any scenario matching degree is higher than a preset threshold.

[0017] The tagging module tags the text in the SMS message to be sent that corresponds to the identified semantics.

[0018] The filtering module reads unlabeled text. If the number of characters in the unlabeled text exceeds a preset character threshold, it obtains the characters with similar shapes and similar pronunciations for each character in the unlabeled text.

[0019] The second recognition module attempts to perform semantic recognition based on the combination of similar-looking and similar-sounding characters in the unlabeled text, and obtains several semantic recognition results.

[0020] The comparison module compares the identified semantic results with a number of preset prohibited scenarios to obtain a second scenario matching degree;

[0021] The third interception module intercepts the SMS message to be sent if any second scenario matching degree is higher than a preset threshold.

[0022] Thirdly, embodiments of this specification provide an electronic device, including a processor and a memory;

[0023] The processor is connected to the memory;

[0024] The memory is used to store executable program code;

[0025] The processor runs a program corresponding to the executable program code stored in the memory to perform the method described in any of the above aspects.

[0026] Fourthly, embodiments of this specification provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the methods described in any of the above aspects.

[0027] Fifthly, embodiments of this specification provide a computer program product, including a computer program that, when executed by a processor, implements the methods described in any of the above aspects.

[0028] The beneficial effects of the technical solutions provided in some embodiments of this specification include at least the following:

[0029] In several embodiments of this specification, the intelligent SMS platform management method and system, by combining keyword filtering, semantic recognition, and character shape and pronunciation similarity detection, can effectively identify variant words (such as homophones and similar-looking characters) and metaphorical violations. This effectively improves the accuracy of identifying illegal SMS messages and reduces false alarms and missed alarms. A method for dynamically generating prohibited scenarios is adopted; that is, through learning and clustering analysis of historical illegal SMS messages, scenario descriptions corresponding to different violation types are automatically generated, and the risk control model is continuously updated and improved accordingly, enabling it to adapt to new types of violations.

[0030] Other features and advantages of various embodiments of this specification will be further revealed in the following detailed description and accompanying drawings. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments of this specification, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 This is a schematic diagram of the intelligent SMS platform management method provided in the embodiments of this specification.

[0033] Figure 2 This is a schematic diagram of the intelligent SMS platform management method provided in the embodiments of this specification.

[0034] Figure 3 This is a schematic diagram illustrating the correlation of scenarios described in the embodiments of this specification.

[0035] Figure 4 This is a schematic diagram of the intelligent SMS platform management system provided in the embodiments of this specification.

[0036] Figure 5 A schematic diagram of an electronic device provided in an embodiment of this specification. Detailed Implementation

[0037] The technical solutions of the embodiments of this specification will be explained and described below with reference to the accompanying drawings. However, the following embodiments are only preferred embodiments of this specification and not all of them. Other embodiments obtained by those skilled in the art based on the embodiments in the implementation methods without creative effort are all within the protection scope of this specification.

[0038] The terms "first," "second," "third," etc., in the description, claims, and accompanying drawings are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such processes, methods, products, or apparatus.

[0039] In the following description, terms such as “inner,” “outer,” “upper,” “lower,” “left,” and “right” are used only to facilitate the description of the embodiments and to simplify the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this specification.

[0040] All data involved in this application are information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0041] Before introducing the technical solutions described in this manual, the application scenarios and related technologies of the technical solutions will be introduced.

[0042] SMS has become a widely used communication method, offering great convenience. However, this convenience also brings security risks and management challenges, such as spam, fraudulent messages, leaks of sensitive information, and the spread of illegal content. The proliferation of meaningless advertising messages negatively impacts user experience. SMS messages masquerading as legitimate organizations and containing false information may also be involved in fraudulent activities, such as phishing and fake notifications. The malicious dissemination of unauthorized personal information via SMS can lead to privacy breaches. Furthermore, SMS platforms can be used to spread illegal information. Effective risk management measures can not only protect users from various forms of harassment and fraud but also help businesses avoid legal risks arising from improper use. Improving the overall quality of SMS services and user satisfaction is crucial for promoting healthy business development.

[0043] For example, third-party SMS platforms are operated by independent service providers, enabling businesses or individuals to send SMS messages. They typically offer API interfaces, allowing users to send SMS messages in bulk via code, suitable for various scenarios such as verification code sending, marketing promotions, and notification reminders. To avoid the risks associated with SMS, risk management is crucial for third-party SMS platforms. Risk management prevents these platforms from sending mass SMS messages containing illegal content or meaningless advertising, thus improving user experience and creating a cleaner SMS environment.

[0044] To ensure that text messages comply with regulations without causing inconvenience to users, please refer to the appendix. Figure 1 When using a multi-level risk control-based intelligent SMS platform management system, the system first reviews the incoming SMS text (10) by comparing it against a database of blocked keywords (21) to identify potentially illegal content. Next, it analyzes the semantics (22) of the SMS. Based on the semantics, it determines whether prohibited scenarios (23) exist, such as whether it contains misleading information, false advertising, or other harmful content. If the SMS passes the initial review, it further checks for text that does not correspond to the identified semantics and attempts to identify more potentially hidden risk information. This significantly improves SMS compliance and protects users from illegal SMS messages.

[0045] This manual first provides a management method for an intelligent SMS platform based on multi-level risk control. Please refer to the appendix. Figure 2 The steps include:

[0046] Step S1) Read the text message 10 to be sent and compare it with a preset interception keyword database 21. If a matching interception keyword 21 is found, the text message to be sent is intercepted. The text messages 10 are compared with a preset interception keyword database 21. The interception keyword database 21 typically contains words and phrases related to illegal, sensitive, or inappropriate content. Examples include fraud-related terms, illegal advertising language, and prohibited product sales terms. For instance, phrases like "prize notification," "abnormal bank account," and "click the link to claim a reward" are included in the interception keyword database 21. If any keyword matching the interception keyword database 21 is found in the text message 10, the text message will be intercepted. This process can initially filter text messages containing obvious violations.

[0047] Step S2) Identify the semantics 22 of the SMS message to be sent, compare the semantics with a number of preset prohibited scenarios 23, and obtain the scenario matching degree.

[0048] The semantics 22 of the text message to be sent is identified using natural language processing technology, which can capture the implicit meaning and sentiment in the text. For example, the semantics 22 of a promotional text message such as "All items are 90% off, miss it today or you'll have to wait another year" can be interpreted as the text message promoting a limited-time discount event with a discount of 90%. This also includes labeling and classifying the text message content. Machine learning models are used to determine the text message's subject matter (e.g., advertisement, notification, warning), its specific emotional tone (e.g., positive, negative, neutral), and its degree of indifference (e.g., moderate, aggressive).

[0049] The methods for pre-setting prohibited scenarios 23 include:

[0050] Read historical illegal text messages, identify the semantics 22 of the illegal text messages, and cluster them according to the semantics to obtain illegal text message clusters;

[0051] Receive manually labeled violation information for each cluster of violation-related SMS messages;

[0052] Based on the illegal SMS messages included in the aforementioned illegal SMS message cluster, a scenario description corresponding to the illegal type is generated;

[0053] Based on the scenario description and the violation type, prohibited scenario 23 is obtained.

[0054] Collect similar past cases of fraudulent text messages. For example: "Dear customer, your account has an anomaly. Please click the following link immediately to complete verification: [Malicious Link]". Analyze the semantics of these text messages using natural language processing technology.22 The main intention of this text message is to lead the recipient to believe that there is a problem with their bank account and induce them to click the provided link. Group text messages with similar intentions and patterns into one category. All text messages containing "account anomaly," "please verify immediately," and providing unknown links will be grouped into the same fraudulent text message cluster. Common characteristics of attacks used to subsequently identify fraudulent text message clusters. Each category of fraudulent text messages is labeled, such as the aforementioned text message being labeled "phishing".

[0055] Based on these classifications and tags, detailed scenario descriptions are created. For example, "The text message claims that the user's bank account has an anomaly and provides an external link for the user to verify the account" is used as a scenario description, and after being associated with the violation type of "phishing," it is designated as prohibited scenario 23.

[0056] For example, the text message to be sent might read: "Hello, abnormal transactions have been detected on your bank card. Please log in to the following website as soon as possible to confirm the details: [Suspicious Link]". The semantics of the text message to be sent, identified as 22, are: This text message conveys the information that the user's bank card may have had abnormal transactions, and provides a link for the recipient to click and view.

[0057] Compared with several pre-defined prohibited scenarios 23, the semantics 22 of this text message are compared with the previously defined prohibited scenario 23 of the "phishing" type. The text message content conforms to the pattern of "claiming an abnormal account and providing an external link," and is therefore determined to be a phishing situation.

[0058] On the other hand, the method for comparing the semantics with a number of preset prohibited scenarios 23 to obtain the scenario matching degree includes:

[0059] The semantics are compared with the scene description of each prohibited scene 23 to obtain the correlation between the semantics and the scene description;

[0060] The scene matching degree is obtained based on the correlation.

[0061] When obtaining the scenario matching score, the relevance of the text message to be sent, "Hello, an abnormal transaction has been detected on your bank card. Please log in to the following website as soon as possible to confirm the details: [suspicious link]", to the scenario description of the prohibited scenario 23, "phishing". For example, the relevance score is obtained by identifying and comparing keywords, phrases, and the overall semantic structure. This can be done using natural language processing models disclosed in the art.

[0062] On the other hand, this specification provides an improved method for obtaining the correlation between the semantics and the scene description, specifically including the following steps:

[0063] Receive several value descriptions for each prohibited scenario 23;

[0064] Several scene sub-descriptions 25 are generated based on each value description, and the several scene sub-descriptions 25 constitute the scene description;

[0065] The semantics are compared with the scene sub-description 25 respectively to obtain the matching degree between the semantics and the scene sub-description 25;

[0066] The maximum value of the matching degree between the semantics and the scene sub-description 25 is taken as the correlation degree between the semantics and the scene description.

[0067] For example, the value description of the prohibited scenario 23 for "phishing" is "able to eliminate account risks" and "clicking the link allows for convenient processing". Based on the value description, scenario sub-description 25 is generated.

[0068] The sub-description 25 for the scenario "able to eliminate account risks" is: "The text message claims that the recipient's account is at risk or abnormal, such as account theft or unauthorized operation, and implies that if no action is taken (usually by clicking a link), it will lead to further security problems. Such text messages usually attempt to induce panic in the recipient, making them believe that only by following the instructions can potential losses be avoided."

[0069] The sub-description of scenario 25 corresponding to "clicking the link can facilitate processing" is: "The text message encourages the recipient to solve a problem or complete a task by clicking the provided link, claiming that this method is the most convenient and effective solution. The link may point to a webpage that appears legitimate but is actually used to collect personal information or carry out other illegal activities."

[0070] Please see the appendix Figure 3 For each scenario sub-description 25, the system analyzes whether the SMS text 10 contains related keywords, phrases, and overall semantic structure. For example, if the SMS mentions "account security threat" or "access the following link immediately," it is highly relevant to the first scenario sub-description 25. If the SMS mentions "click the link to verify identity," it is highly relevant to the second scenario sub-description 25.

[0071] The matching degree between the text message and each scenario sub-description 25 was calculated. The text message to be sent, "Urgent Notice: Your bank account may be under security threat. Please visit the following link immediately to protect your funds." has a very high matching degree with the first scenario sub-description 25, with a matching degree of 80%, and also has some relevance with the second scenario sub-description 25, with a matching degree of 60%.

[0072] Ultimately, the highest matching score was selected as the correlation score between the text message and the entire "phishing" prohibited scenario 23, which is 80%.

[0073] Step S3) If any scenario matching degree is higher than the preset threshold, the SMS message to be sent is intercepted.

[0074] If the matching rate exceeds the set threshold (e.g., 70%), the SMS message to be sent is deemed risky and will be blocked.

[0075] Step S4) Based on the identified semantics, mark the text in the SMS message to be sent that corresponds to the semantics.

[0076] Once the semantic analysis 22 of the text message 10 to be sent is completed, the semantics 22 of the text message to be sent are determined. For example, if the text message to be sent involves false advertising, texts about "limited-time discounts" and "miss today and wait another year" will be identified. Based on the identified semantics 22, the parts of the text message 10 directly related to these semantics are found. For example, in the text message "All items are 90% off, miss today and wait another year," "All items are 90% off" and "miss today and wait another year" will be marked as semantically related text segments. After marking, the next step is to read the unmarked text parts (if any) and perform additional security checks on these parts.

[0077] Another exemplary text message to be sent is "Dear trainee [online], thank you [on] for choosing the Smart Learning Platform! [Bo] Sign up now and enjoy an 80% discount for a limited time! [Color] For details, please visit: [suspicious link]. Wish you a pleasant learning experience!". Most of the content is normal and compliant, but the content within the brackets describes the true purpose of the suspicious link as online illegal content. This is used to evade the security checks of the text message platform and at the same time attract people who want to participate in illegal activities to click on the suspicious link.

[0078] Step S5) Read the unmarked text. If the number of characters in the unmarked text exceeds the preset character count threshold, obtain the characters with similar glyphs and similar pronunciations for each character in the unmarked text.

[0079] Exemplarily, when the text message to be sent is "Dear trainee [limit], thank you [still] for choosing the Smart Learning Platform! [Thin] Sign up now and enjoy an 80% discount for a limited time! [Pick] For details, please visit: [suspicious link]. Wish you a pleasant learning experience!". This will further increase the difficulty of the text message security monitoring to identify and intercept.

[0080] Among them, the method for obtaining the characters with similar glyphs and similar pronunciations for each character in the unmarked text includes:

[0081] Receive the common character library and group the common character library according to similar glyphs and similar pronunciations respectively;

[0082] Read a number of pre-configured sample corpus texts, and randomly replace several characters in the sample corpus texts with characters with similar glyphs or similar pronunciations in the same group;

[0083] Receive the replaced characters found by the testers after reading at the preset reading speed, and calculate the probability of the replacement being discovered;

[0084] Take the characters with similar glyphs whose probability is less than the preset reference probability value as the characters with similar glyphs of the replaced characters, and take the characters with similar pronunciations whose probability is less than the preset reference probability value as the characters with similar pronunciations of the replaced characters.

[0085] For example, use the 3,500 common characters in the "List of Commonly Used Characters in Modern Chinese" and group them according to similar glyphs and similar pronunciations. For similar glyphs: such as "wood" and "technique", "sun" and "yue", "person" and "enter". For similar pronunciations: such as "four" and "ten", "is" and "matter", "walk" and "shape".

[0086] Extract several sample corpora from real text messages or historical data. Randomly replace some characters in the corpus with characters that are similar in shape or pronunciation. For example: original sentence: "Click on the link to receive the reward", after replacement: "Click on the link to receive the award assistance". Provide the replaced corpus to the testers and let them read it at a preset reading speed. Record which characters are recognized and which are not.

[0087] For each replaced character, count the proportion of times it is detected. If the probability of a character being detected after replacement < preset reference probability value (such as 30%), then this character is considered similar enough to the original character and can be added to the corresponding character library as a "character with similar shape" or "character with similar pronunciation". Finally, form a character library of characters with similar shapes and a character library of characters with similar pronunciations.

[0088] The system reads the unmarked text in the text message and searches for candidate replacement characters for each character in the "character library of characters with similar shapes" or "character library of characters with similar pronunciations". For example: when the character "shang" is found, check if it is a character with a similar shape to "shang". If so, try to restore "shang" to "shang" and re-analyze the semantics.

[0089] Step S6) Try to perform semantic recognition based on the combination of the characters with similar shapes and characters with similar pronunciations in the unmarked text, and obtain several recognition semantic results 24.

[0090] The method of trying to perform semantic recognition based on the combination of the characters with similar shapes and characters with similar pronunciations in the unmarked text and obtaining several recognition semantic results 24 includes:

[0091] Read the characters of the unmarked text in sequence as the inspected characters, and obtain all the characters with similar shapes and characters with similar pronunciations of the character as the inspection set;

[0092] Respectively obtain the appearance probabilities of the possible subsequent characters of the characters in the inspection set, and select the possible subsequent characters with appearance probabilities higher than the preset value;

[0093] Read the next character of the inspected character in the unmarked text. When the possible subsequent character belongs to the characters with similar shapes or characters with similar pronunciations of the next character, take the next character and the inspected character as a combination. If the combination includes the same inspected character, merge the combinations;

[0094] After all the characters of the unmarked text have been used as inspected characters, obtain all the combinations;

[0095] Try to recognize the semantics of the combination 22, and obtain the recognition semantic result 24 according to the successfully recognized semantics 22.

[0096] The unmarked text includes "[Limit]", "[Shang]", "[Thin]", "[Collect]".

[0097] The characters with similar glyphs to "限" are "眼, 根, 恨, 艰", and the characters with similar pronunciations to "限" are "现, 线, 宪, 闲, 县".

[0098] The characters with similar glyphs to "尚" are "常, 当, 尝, 赏", and the characters with similar pronunciations to "尚" are "商, 上, 伤, 赏, 绍".

[0099] Regarding the character with a similar glyph to "限", which is "眼". The subsequent possible characters and their occurrence probabilities for "眼" are: 睛 (30%), 光 (35%), 神 (20%), 镜 (20%), 科 (20%), 泪 (25%).

[0100] Regarding the character with a similar pronunciation to "限", which is "线". The subsequent possible characters and their occurrence probabilities for "线" are: 条 (5%), 缆 (10%), 上 (35%), 下 (35%), 路 (15%), 索 (5%).

[0101] It can be seen that the occurrence probability of the subsequent possible character "上" for the character with a similar pronunciation to "限", which is "线", is 35%, and it exists among the characters with similar glyphs and similar pronunciations of the next unmarked character "

尚

[0102] Step S7) Compare the recognized semantic result 24 with a number of preset prohibited scenarios 23 to obtain a second scenario matching degree.

[0103] Attempt to perform semantic recognition on "线上博彩" to obtain a recognized semantic result 24, and compare the recognized semantic result 24 with a number of preset prohibited scenarios 23, and a relatively high second scenario matching degree with "网络钓鱼" can be obtained.

[0104] Step S8) If there is any second scenario matching degree higher than the preset threshold, then intercept the to-be-sent short message.

[0105] When the second scenario matching degree with "网络钓鱼" is relatively high (higher than the preset threshold), intercept the to-be-sent short message. To achieve more perfect short message risk control.

[0106] On the other hand, this specification provides an intelligent short message platform control system based on multi-level risk control. Please refer to the appendix Figure 4 , including:

[0107] The first interception module 100 reads the to-be-sent short message text 10, compares the short message text 10 with a preset interception keyword 21 library. If there is a matching interception keyword 21, then intercept the to-be-sent short message;

[0108] The first identification module 200 identifies the semantics 22 of the SMS message to be sent, compares the semantics with a number of preset prohibited scenarios 23, and obtains the scenario matching degree.

[0109] The second interception module 300 intercepts the SMS message to be sent if any scenario matching degree is higher than a preset threshold.

[0110] The tagging module 400 tags the text in the SMS message to be sent that corresponds to the identified semantics.

[0111] The filtering module 500 reads unmarked text. If the number of characters in the unmarked text exceeds a preset character threshold, it obtains characters with similar shapes and similar pronunciations for each character in the unmarked text.

[0112] The second recognition module 600 attempts to perform semantic recognition based on the combination of similar-looking characters and similar-sounding characters in the unmarked text, and obtains several semantic recognition results 24.

[0113] The comparison module 700 compares the semantic recognition result 24 with a number of preset prohibited scenarios 23 to obtain a second scenario matching degree.

[0114] The third interception module 800 intercepts the SMS message to be sent if any second scenario matching degree is higher than a preset threshold.

[0115] Please see Figure 5 The diagram shown is a structural schematic of an electronic device provided in an embodiment of this specification.

[0116] like Figure 5As shown, the electronic device 1100 may include: at least one processor 1101, at least one network interface 1104, a user interface 1103, a memory 1105, and at least one communication bus 1102. The communication bus 1102 can be used to connect and communicate with the various components mentioned above. The user interface 1103 may include buttons, and optionally may include standard wired or wireless interfaces. The network interface 1104 may include, but is not limited to, a Bluetooth module, an NFC module, or a Wi-Fi module. The processor 1101 may include one or more processing cores. The processor 1101 connects to various parts within the electronic device 1100 using various interfaces and lines, and performs various functions of the routing device and processes data by running or executing instructions, programs, code sets, or instruction sets stored in the memory 1105, and by calling data stored in the memory 1105. Optionally, the processor 1101 may be implemented using at least one hardware form of DSP, FPGA, or PLA. The processor 1101 may integrate one or more combinations of CPU, GPU, and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content that the display screen needs to show; and the modem is used for wireless communication.

[0117] It is understandable that the aforementioned modem may not be integrated into the processor 1101, but may be implemented using a separate chip.

[0118] The memory 1105 may include RAM or ROM. Optionally, the memory 1105 may include a non-transitory computer-readable medium. The memory 1105 may be used to store instructions, programs, code, code sets, or instruction sets. The memory 1105 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. Optionally, the memory 1105 may also be at least one storage device located remotely from the aforementioned processor 1101. As a computer storage medium, the memory 1105 may include an operating system, a network communication module, a user interface module, and application programs. The processor 1101 may be used to call the application programs stored in the memory 1105 and execute the methods in the above-described embodiments.

[0119] This specification also provides a computer-readable storage medium storing instructions that, when executed on a computer or processor, cause the computer or processor to perform multiple steps as described in the above embodiments. If the constituent modules of the above-described electronic device are implemented as software functional units and sold or used as independent products, they can be stored in the computer-readable storage medium.

[0120] This specification also provides a computer program product, including a computer program that, when executed by a processor, implements the multiple steps described in the above embodiments.

[0121] Where there is no conflict, the technical features in this embodiment and implementation scheme can be combined arbitrarily.

[0122] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented, in whole or in part, as a computer program product. The computer program product includes multiple computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this specification are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted through the computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center integrating multiple available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., digital versatile discs (DVDs)), or semiconductor media (e.g., solid-state drives (SSDs)).

[0123] When implemented through hardware or firmware, the aforementioned method flow is programmed into the hardware circuit to obtain the corresponding hardware circuit structure and achieve the corresponding function. For example, a Programmable Logic Device (PLD) (such as a Field Programmable Gate Array (FPGA)) is such an integrated circuit, whose logic function is determined by the user programming the device. Designers can program a digital system onto a PLD themselves, eliminating the need for chip manufacturers to design and fabricate dedicated integrated circuit chips. Furthermore, nowadays, instead of manually fabricating integrated circuit chips, this programming is mostly implemented using "logic compiler" software, similar to the software compiler used in program development. The original code before compilation must also be written in a specific programming language, called a Hardware Description Language (HDL). There is not just one HDL, but many. Those skilled in the art should understand that by simply performing some logic programming on the method flow using one of the aforementioned hardware description languages ​​and programming it into an integrated circuit, the hardware circuit implementing the logic method flow can be easily obtained.

[0124] The embodiments described above are merely preferred embodiments of this specification and are not intended to limit the scope of this specification. Any modifications and improvements made by those skilled in the art to the technical solutions of this specification without departing from the spirit of this specification should fall within the protection scope defined by the claims of this specification.

Claims

1. A method for managing an intelligent SMS platform based on multi-level risk control, characterized in that, Including the following steps: Read the text message to be sent, compare the text message with a preset block keyword library, and if a matching block keyword is found, block the text message to be sent. Identify the semantics of the SMS message to be sent, compare the semantics with several preset prohibited scenarios, and obtain the scenario matching degree; If any scenario has a matching degree higher than a preset threshold, the SMS message to be sent will be blocked. Based on the identified semantics, mark the text in the SMS message to be sent that corresponds to the semantics; Read unmarked text. If the number of characters in the unmarked text exceeds a preset character threshold, obtain the characters with similar shapes and similar pronunciations for each character in the unmarked text. Based on the combination of similar-looking and similar-sounding characters in the unlabeled text, semantic recognition is attempted to obtain several semantic recognition results; The second scene matching degree is obtained by comparing the semantic recognition result with a number of preset prohibited scenes; If any second scenario has a matching degree higher than a preset threshold, the SMS message to be sent will be intercepted.

2. The intelligent SMS platform management and control method based on multi-level risk control according to claim 1, characterized in that, Methods for pre-setting prohibited scenarios include: Read historical illegal text messages, identify the semantics of the illegal text messages, and cluster them according to the semantics to obtain illegal text message clusters; Receive manually labeled violation information for each cluster of violation-related SMS messages; Based on the illegal SMS messages included in the aforementioned illegal SMS message cluster, a scenario description corresponding to the illegal type is generated; Based on the scenario description and the violation type, the prohibited scenario is obtained.

3. The intelligent SMS platform management and control method based on multi-level risk control according to claim 2, characterized in that, The method for comparing the semantics with a set of preset prohibited scenarios to obtain the scenario matching degree includes: The semantics are compared with the scene description of each prohibited scenario to obtain the correlation between the semantics and the scene description; The scene matching degree is obtained based on the correlation.

4. The intelligent SMS platform management and control method based on multi-level risk control according to claim 3, characterized in that, Methods for obtaining the correlation between the semantics and the scene description include: Receive several value descriptions for each prohibited scenario; Several scene sub-descriptions are generated based on each value description, and the several scene sub-descriptions constitute the scene description; The semantics are compared with the scene sub-descriptions respectively to obtain the matching degree between the semantics and the scene sub-descriptions; The maximum matching degree between the semantics and the scene sub-description is taken as the correlation degree between the semantics and the scene description.

5. The intelligent SMS platform management and control method based on multi-level risk control according to claim 1, characterized in that, Methods for obtaining glyphically similar and phonetically similar characters for each character in untagged text include: Receive a commonly used character set and group the commonly used character sets according to similar character shapes and similar pronunciations; Read several pre-configured sample texts and randomly replace several characters in the sample texts with characters with similar shapes or similar pronunciations in the same group; The system receives replaced characters found by testers after reading at a preset reading speed, and calculates the probability that the replacement was detected. Characters with similar shapes whose probability is less than a preset reference probability value are considered as similar characters in shape to the replaced character, and characters with similar pronunciation whose probability is less than a preset reference probability value are considered as similar characters in pronunciation to the replaced character.

6. The intelligent SMS platform management and control method based on multi-level risk control according to claim 1, characterized in that, Based on the combination of similar-looking and similar-sounding characters in the unlabeled text, a method is used to attempt semantic recognition and obtain several semantic recognition results, including: Read the characters of the unlabeled text in sequence as the characters to be examined, and obtain all characters with similar shapes and similar pronunciations as the examination set; The probability of the next possible character in the set of characters is obtained, and the possible characters with a probability higher than a preset value are selected. Read the next character of the examined character in the unmarked text. When the possible character belongs to the character with similar shape or similar pronunciation to the next character, combine the next character and the examined character as a group. If the group includes the same examined character, merge the groups. All characters in the unmarked text were used as examination characters to obtain all combinations; Attempt to identify the semantics of the combination, and obtain the semantic identification result based on the successfully identified semantics.

7. A smart SMS platform management and control system based on multi-level risk control, characterized in that, include: The first interception module reads the text of the SMS to be sent, compares the text of the SMS with a preset interception keyword library, and if a matching interception keyword is found, then the SMS to be sent is intercepted. The first identification module identifies the semantics of the SMS message to be sent, compares the semantics with a number of preset prohibited scenarios, and obtains the scenario matching degree. The second interception module intercepts the SMS message to be sent if any scenario matching degree is higher than a preset threshold. The tagging module tags the text in the SMS message to be sent that corresponds to the identified semantics. The filtering module reads unlabeled text. If the number of characters in the unlabeled text exceeds a preset character threshold, it obtains the characters with similar shapes and similar pronunciations for each character in the unlabeled text. The second recognition module attempts to perform semantic recognition based on the combination of similar-looking and similar-sounding characters in the unlabeled text, and obtains several semantic recognition results. The comparison module compares the identified semantic results with a number of preset prohibited scenarios to obtain a second scenario matching degree; The third interception module intercepts the SMS message to be sent if any second scenario matching degree is higher than a preset threshold.

8. An electronic device, characterized in that, Including the processor and memory; The processor is connected to the memory; The memory is used to store executable program code; The processor runs a program corresponding to the executable program code stored in the memory to perform the method as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-6.

Images