CN118035808A - News release method, news release device, electronic equipment and storage medium - Google Patents
News release method, news release device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN118035808A CN118035808A CN202311837593.2A CN202311837593A CN118035808A CN 118035808 A CN118035808 A CN 118035808A CN 202311837593 A CN202311837593 A CN 202311837593A CN 118035808 A CN118035808 A CN 118035808A
- Authority
- CN
- China
- Prior art keywords
- news
- distributed
- false
- condition
- under
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 60
- 238000003860 storage Methods 0.000 title claims abstract description 14
- 238000001514 detection method Methods 0.000 claims abstract description 152
- 238000012549 training Methods 0.000 claims abstract description 11
- 238000009826 distribution Methods 0.000 claims description 25
- 238000004590 computer program Methods 0.000 claims description 13
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 238000004891 communication Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000008451 emotion Effects 0.000 description 4
- 238000004140 cleaning Methods 0.000 description 2
- 238000012217 deletion Methods 0.000 description 2
- 230000037430 deletion Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000037361 pathway Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000007480 spreading Effects 0.000 description 1
- 238000003892 spreading Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a news publishing method, a news publishing device, electronic equipment and a storage medium, and relates to the technical field of artificial intelligence, wherein the method comprises the following steps: acquiring news to be distributed; inputting the news to be distributed into a news detection model to obtain a detection result output by the news detection model, wherein the detection result is used for representing the true and false conditions of the news to be distributed, and the news detection model is obtained by training an initial news detection model based on sample news and label information used for representing the true and false conditions of the sample news; and publishing the news to be published under the condition that the detection result represents that the news to be published is real news. When the news to be released is acquired, the true and false detection is carried out on the news to be released based on the news detection model, and release is carried out only under the condition that the news to be released is true news, so that the accuracy of news release is improved.
Description
Technical Field
The present invention relates to the field of artificial intelligence technologies, and in particular, to a news publishing method, a news publishing device, an electronic device, and a storage medium.
Background
News, also called messages or information, is a name for information disseminated through media pathways such as newspapers, radio, television stations, and the like. Is a kind of cultural relics for recording society, spreading information and reflecting age.
In the related art, when receiving news input by a user, a news publishing platform directly publishes the news input by the user.
However, in the related art, the false detection is not performed on the news input by the user, so that false news may exist in the released news, thereby reducing the reliability of news release.
Disclosure of Invention
Aiming at the problems existing in the prior art, the invention provides a news publishing method, a news publishing device, electronic equipment and a storage medium.
The invention provides a news publishing method, which comprises the following steps:
Acquiring news to be distributed;
Inputting the news to be distributed into a news detection model to obtain a detection result output by the news detection model, wherein the detection result is used for representing the true and false conditions of the news to be distributed, and the news detection model is obtained by training an initial news detection model based on sample news and label information used for representing the true and false conditions of the sample news;
And publishing the news to be published under the condition that the detection result represents that the news to be published is real news.
The invention also provides a news publishing method, which further comprises the following steps:
Under the condition that the detection result represents that the news to be distributed is false news, prohibiting the news to be distributed from being distributed, and acquiring a target user identification for inputting the news to be distributed;
acquiring the quantity of false news input by a user represented by the target user identifier in a preset history time;
And adding the target user identification into a blacklist under the condition that the number is larger than a first preset number.
The invention also provides a news publishing method, which is used for acquiring the quantity of false news input by the user represented by the target user identifier in a preset history time length, and comprises the following steps:
determining the news type of the news to be distributed;
Acquiring the number of false news corresponding to the news type input by the user in a preset history time;
and adding the target user identifier to a blacklist under the condition that the number is larger than a first preset number, wherein the method comprises the following steps:
And adding the target user identifier into a blacklist under the condition that the number of false news corresponding to the news type is larger than the first preset number corresponding to the news type.
The invention also provides a news publishing method, which further comprises the following steps:
Acquiring real reference news which has similar semantics to the news headlines of the news to be distributed and has similar semantics to news contents under the condition that the number is smaller than or equal to a first preset number;
And outputting prompt information, wherein the prompt information comprises real reference news, and the prompt information is used for reminding the user to modify the news to be distributed based on the real reference news.
The invention also provides a news publishing method, which further comprises the following steps of:
Acquiring comment information of the news to be distributed;
And under the condition that the number of comment information representing that the news to be distributed is false news is larger than a second preset number, updating the news detection model based on the news to be distributed and label information representing that the news to be distributed is false news.
The invention also provides a news publishing method, which publishes the news to be published under the condition that the detection result represents that the news to be published is real news, and comprises the following steps:
Determining whether the news to be distributed accords with a distribution rule under the condition that the detection result represents that the news to be distributed is real news;
determining whether the news to be distributed contains sensitive words or not based on a sensitive word stock under the condition that the news to be distributed accords with the distribution rule;
and publishing the news to be published under the condition that the news to be published does not contain sensitive words.
The invention also provides a news publishing method, which further comprises the following steps:
Acquiring context information containing sensitive words under the condition that the news to be distributed contains the sensitive words;
Determining whether the sensitive word is related to semantic information of the context information;
Under the condition that the sensitive word is related to semantic information of the context information, determining a target replacement word corresponding to the sensitive word based on a word replacement library, replacing the sensitive word with the target replacement word to obtain replaced news to be distributed, and distributing the replaced news to be distributed;
And deleting the sensitive words under the condition that the sensitive words are irrelevant to the semantic information of the context information, obtaining deleted news to be distributed, and distributing the deleted news to be distributed.
The invention also provides a news publishing device, which comprises:
the first acquisition unit is used for acquiring news to be distributed;
the detection unit is used for inputting the news to be distributed into a news detection model to obtain a detection result output by the news detection model, wherein the detection result is used for representing the true and false conditions of the news to be distributed, and the news detection model is obtained by training an initial news detection model based on sample news and label information for representing the true and false conditions of the sample news;
the first issuing unit is used for issuing the news to be issued under the condition that the detection result represents that the news to be issued is real news.
The invention also provides an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing a news distribution method as described above when executing the program.
The present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a news distribution method as described in any of the above.
The invention also provides a computer program product comprising a computer program which when executed by a processor implements a news distribution method as described in any one of the above.
According to the news publishing method, the device, the electronic equipment and the storage medium, when the news to be published is acquired, the news to be published is input into the news detection model, the detection result output by the news detection model is obtained, and when the detection result represents that the news to be published is real news, the news to be published is published. According to the news publishing method and device, when the news to be published is acquired, the true and false detection is carried out on the news to be published based on the news detection model, and the news to be published is published only under the condition that the news to be published is the true news, so that the accuracy of news publishing is improved.
Drawings
In order to more clearly illustrate the invention or the technical solutions of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic flow diagram of a news publishing method according to an embodiment of the present invention;
FIG. 2 is a second flowchart of a news publishing method according to an embodiment of the present invention;
FIG. 3 is a third flow chart of a news publishing method according to an embodiment of the present invention;
FIG. 4 is a flowchart of a news publishing method according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of a news publishing device according to an embodiment of the present invention;
fig. 6 is a schematic diagram of an entity structure of an electronic device according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
False news is widely spread on networks and social media as a false and misleading news content, and brings negative effects to society and individuals. At present, detection and recognition technology for false news becomes an important research field, and real news and false news need to be rapidly and accurately distinguished so as to ensure that the public obtains accurate information.
The news publishing method of the present invention is described below with reference to fig. 1 to 4. The execution subject of the news publishing method may be an electronic device such as a terminal or a computer, or may be a news publishing device provided in the electronic device, and the news publishing device may be implemented by software, hardware, or a combination of both.
Fig. 1 is a schematic flow chart of a news publishing method according to an embodiment of the present invention, as shown in fig. 1, the news publishing method includes the following steps:
Step 101, acquiring news to be distributed.
When the user needs to release news, the news release instruction can be clicked on the news release platform of the electronic device, the news input interface is output when the electronic device receives the news release instruction, and the user can input news to be released on the news input interface, so that the electronic device obtains the news to be released.
Step 102, inputting the news to be distributed into a news detection model to obtain a detection result output by the news detection model, wherein the detection result is used for representing the true and false situation of the news to be distributed, and the news detection model is obtained by training an initial news detection model based on sample news and label information used for representing the true and false situation of the sample news.
When the news to be distributed is obtained, the news to be distributed is subjected to preprocessing operations such as cleaning, word segmentation and stop word removal to obtain preprocessed news to be distributed, the preprocessed news to be distributed is converted into the news to be distributed in a vector format which can be identified by a news detection model, the news to be distributed in the vector format is input into the news detection model, a detection result output by the news detection model is obtained, the detection result is used for representing the true and false situation of the news to be distributed, for example, when the detection result is 1, the news to be distributed is represented as real news, and when the detection result is 0, the news to be distributed is represented as false news.
It should be noted that the news detection model is specifically trained based on the following manner: firstly, collecting a plurality of sample news, performing pretreatment operations such as cleaning, word segmentation and stop word removal on the sample news to obtain pretreated sample news, extracting characteristics of the pretreated sample news, converting the pretreated sample news into sample news in a vector format which can be identified by an initial news detection model according to a forward propagation process, inputting the sample news in the vector format into the initial news detection model, and classifying true and false news of the sample news in the vector format through a classifier of the initial news detection model to obtain a prediction detection result; and constructing a loss function based on the true and false conditions of the sample news represented by the prediction detection result and the labels corresponding to the sample news, optimizing the model parameters of the initial news detection model based on the loss function, and repeatedly iterating and optimizing the initial news detection model to improve the prediction accuracy and generalization capability of the news detection model finally obtained by training.
It should be noted that the detection result may also be a multi-classification result, when the detection result is 1, the news to be distributed is represented as real news, when the detection result is 0, the news to be distributed is represented as false news, and when the detection result is 2, the news to be distributed is represented as whether the real news or the false news is uncertain.
It should be noted that, when the sample news is real news, the label of the sample news is a real news label, and when the sample news is false news, the label of the sample news is a false news label. The initial news detection model may be a detection model constructed based on deep learning techniques, such as transfomer or BERT, etc.
And step 103, publishing the news to be published under the condition that the detection result represents that the news to be published is real news.
For example, when the detection result output by the news detection model indicates that the news to be distributed is real news, the news to be distributed is the compliance news, and the news to be distributed can be directly distributed.
According to the news publishing method, when the news to be published is acquired, the news to be published is input into the news detection model, the detection result output by the news detection model is obtained, and when the detection result represents that the news to be published is real news, the news to be published is published. According to the news publishing method and device, when the news to be published is acquired, the true and false detection is carried out on the news to be published based on the news detection model, and the news to be published is published only under the condition that the news to be published is the true news, so that the accuracy of news publishing is improved.
In an embodiment, fig. 2 is a second flowchart of a news publishing method according to an embodiment of the present invention, as shown in fig. 2, the news publishing method further includes the following steps:
step 104, prohibiting the release of the news to be released and acquiring a target user identification for inputting the news to be released under the condition that the detection result represents that the news to be released is false news.
The false news is a public opinion that a deceptive party can release false information to achieve a certain purpose, and cannot truly reflect the original appearance of an objective object, and has a report of false components.
When the detection result output by the news detection model indicates that the news to be distributed is false news, the news to be distributed is not compliant, the news to be distributed is forbidden to be distributed, and the target user identification of the news to be distributed is acquired. The target user identifier may be user account information or other information capable of uniquely characterizing the user.
Step 105, obtaining the quantity of false news input by the user represented by the target user identifier in a preset history time.
The news data list stores the corresponding relation of the user identification, the false news and the input time, when the target user identification is acquired, all the false news corresponding to the target user identification is acquired from the news data list, then the false news in the preset history duration is screened out from all the false news corresponding to the target user identification based on the input time of all the false news corresponding to the target user identification, and the number of the false news in the preset history duration is counted. The preset history period may be set based on the requirement, for example, the preset history period is 3 months.
And 106, adding the target user identification into a blacklist under the condition that the number is larger than a first preset number.
The first preset number may be set based on the requirement, for example, the first preset number is 3.
When the quantity of the false news input in the preset history duration is acquired, the quantity is compared with a first preset quantity, and when the quantity is larger than the first preset quantity, the user represented by the target user identifier is indicated to frequently input the false news, so that the target user identifier is added into a blacklist; and when the number is smaller than or equal to the first preset number, the user represented by the target user identifier is a legal user, and the corresponding relation among the target user identifier, the news to be distributed and the input time is stored in a news data list.
In this embodiment, when the detection result indicates that the news to be distributed is false news, the news to be distributed is forbidden, the accuracy of news distribution is improved, the number of false news input by the user in the preset history duration and represented by the target user identifier of the news to be distributed is acquired, when the number is greater than the first preset number, the user is an illegal user, the target user identifier of the user is added into the blacklist, and the news to be distributed is forbidden directly when the user inputs new news to be distributed again.
In an embodiment, the step 105 obtains the number of false news input by the user represented by the target user identifier in the preset history period, and the step 106 adds the target user identifier to the blacklist if the number is greater than the first preset number, which may be specifically implemented by:
Determining the news type of the news to be distributed; and acquiring the number of false news corresponding to the news type input by the user in a preset history time, and adding the target user identifier into a blacklist under the condition that the number of the false news corresponding to the news type is larger than a first preset number corresponding to the news type.
The news types may be a time type, a social type, an emotion type, a daily life type, an entertainment type and the like, a corresponding first preset number may be preset for each type of news type, for example, news of the time type is serious, the first preset number set for the time type is 1, the first preset number set for the social type is 2, the first preset number set for the emotion type is 3, the first preset number set for the daily life type and the entertainment type is 5, in practical application, the corresponding first preset number may be set based on the priority of the news type, and the higher the priority is, the smaller the corresponding first preset number is; for example, the priority of the administrative type is higher than the priority of the social type, the priority of the social type is higher than the priority of the emotion type, the priority of the emotion type is higher than the priority of the daily life type, and the priority of the entertainment type is the same as the priority of the daily life type.
For example, under the condition that the detection result represents that the news to be distributed is false news, semantic understanding is performed on the news to be distributed, and the news type of the news to be distributed is determined, for example, the news to be distributed is a report on a social topic, the news type of the news to be distributed is the social type, false news which is the same as the news type of the news to be distributed is screened out from all the false news which are input by the user in a preset history time, the number of the false news which is the same as the news type of the news to be distributed is counted, and when the number of the false news which is the same as the news type of the news to be distributed is larger than a first preset number corresponding to the news type, the user is indicated to frequently input the false news of the news type, so that a target user identifier of the user is added into a blacklist.
In this embodiment, a preset number corresponding to each type of news type is preset, the number of false news which is input by the user in a preset history time and is the same as the news type of the news to be distributed is obtained, whether the user is added to the blacklist is determined based on the number of the false news which is the same as the news type of the news to be distributed, and therefore humanization of management of the user who distributes news is improved.
In an embodiment, fig. 3 is a third flowchart of a news publishing method according to an embodiment of the present invention, as shown in fig. 3, the news publishing method further includes the following steps:
And 107, acquiring real reference news which has similar semantics to the news headlines of the news to be distributed and has similar semantics to the news contents under the condition that the number is smaller than or equal to a first preset number.
In an example, under the condition that the number of false news input by a user in a preset history time is smaller than or equal to a first preset number, the user is illustrated as a legal user, at this time, a news headline and news content of news to be distributed are acquired, semantic understanding is carried out on the news headline to obtain a first headline semantic understanding result, semantic understanding is carried out on the news content to obtain a first content semantic understanding result, a second headline semantic understanding result of the news headline of each real news in a news library is determined, the second content semantic understanding result of the news content of each real news in the news library is determined, the first similarity of the first headline semantic understanding result and each second headline semantic understanding result is calculated, the second similarity of the first content semantic understanding result and each second content semantic understanding result is calculated, a first candidate real news set corresponding to the first similarity larger than the first preset value is determined, a second candidate real news set corresponding to the second similarity larger than the second preset value is determined, intersection of the first candidate real news set and the second candidate real news set is determined, and the real news in the intersection of the intersection is determined as real reference news.
And 108, outputting prompt information, wherein the prompt information comprises real reference news, and the prompt information is used for reminding the user to modify the news to be distributed based on the real reference news.
When the real reference news is obtained, the real reference news is displayed, so that a user can modify news to be released based on the displayed real reference news, the modified news to be released is input into a news detection model, and if a new detection result output by the news detection model represents that the modified news to be released is the real news, the modified news to be released is released.
In this embodiment, when the number of false news input by the user in the preset history period is less than or equal to the first preset number, the user is reminded to modify the news to be released based on the acquired real reference news, so as to improve the authenticity of the news.
In an embodiment, fig. 4 is a fourth flowchart of a news publishing method according to an embodiment of the present invention, as shown in fig. 4, after step 103, the news publishing method further includes the following steps:
Step 109, obtaining comment information of the news to be distributed.
For example, after publishing the news to be published, comment information of the news to be published of other users can be obtained, semantic understanding is carried out on all the comment information, and the number of comment information representing that the news to be published is false news is determined; for example, if the comment information is "this news feel is not true", it may be determined that the comment information characterizes the news to be distributed as false news.
Step 110, updating the news detection model based on the news to be distributed and tag information representing that the news to be distributed is false news when the number of comment information representing that the news to be distributed is false news is larger than a second preset number.
The second preset number may be set based on the requirement, for example, the second preset number is 10.
For example, when the number of comment information representing that the news to be distributed is false news is larger than the second preset number, the detection result output by the news detection model is wrong, and at this time, the news detection model is updated based on the news to be distributed and tag information representing that the news to be distributed is false news.
In this embodiment, comment information of news to be distributed is obtained, and based on the news to be distributed and tag information representing that the news to be distributed is false news, a news detection model is updated under the condition that the number of comment information representing that the news to be distributed is larger than a second preset number, so that the updated news detection model can learn more news features, the detection performance of the updated news detection model is improved, and the accuracy of news detection is further improved.
In an embodiment, step 103 may be implemented to publish the news to be published if the detection result indicates that the news to be published is real news, by specifically:
Determining whether the news to be distributed accords with a distribution rule under the condition that the detection result represents that the news to be distributed is real news; determining whether the news to be distributed contains sensitive words or not based on a sensitive word stock under the condition that the news to be distributed accords with the distribution rule; and publishing the news to be published under the condition that the news to be published does not contain sensitive words.
For example, when the detection result represents that the news to be distributed is real news, further determining whether the news to be distributed accords with a preset distribution rule, wherein the distribution rule can be that the news to be distributed needs to include news titles, user signs for inputting the news to be distributed, the number of words of the news to be distributed cannot be lower than a preset number of words, and the like, and when the news to be distributed is determined not to accord with the distribution rule, prohibiting the distribution of the news to be distributed; when the news to be distributed is determined to accord with the distribution rule, further determining whether the news to be distributed contains sensitive words in a sensitive word stock, and when the news to be distributed is determined to not contain the sensitive words, distributing the news to be distributed.
In this embodiment, if the detection result indicates that the news to be distributed is real news, it is further required to check whether the news to be distributed accords with the distribution rule and includes the sensitive word, and only if the news to be distributed accords with the distribution rule and does not include the sensitive word, the news to be distributed is distributed, so that accuracy and readability of the news to be distributed are further improved.
In an embodiment, step 103 may be implemented to publish the news to be published if the detection result indicates that the news to be published is real news, by specifically:
acquiring context information containing sensitive words under the condition that the news to be distributed contains the sensitive words; determining whether the sensitive word is related to semantic information of the context information; under the condition that the sensitive word is related to semantic information of the context information, determining a target replacement word corresponding to the sensitive word based on a word replacement library, replacing the sensitive word with the target replacement word to obtain replaced news to be distributed, and distributing the replaced news to be distributed; and deleting the sensitive words under the condition that the sensitive words are irrelevant to the semantic information of the context information, obtaining deleted news to be distributed, and distributing the deleted news to be distributed.
When determining that the news to be distributed contains sensitive words, acquiring context information of the sensitive words in the news to be distributed, determining whether the sensitive words are related to semantic information of the context information, and when determining that the sensitive words are related to the semantic information of the context information, indicating that the deletion of the sensitive words may cause incomplete semantics of a whole sentence, searching a word library for target replacement words corresponding to the sensitive words, replacing the sensitive words in the news to be distributed with the target replacement words to obtain replaced news to be distributed, and distributing the replaced news to be distributed; for example, if the sensitive word is "nugget", the context information containing the sensitive word is "as early as 2023, the number of nuggets reaches the same, and if the sensitive word is deleted, the context information containing the sensitive word is not completely semantically related, the replacement word corresponding to the sensitive word as" nugget "may be determined as" single young "in the word library based on the correspondence between the sensitive word and the replacement word, and then the" single young "is determined as the target replacement word, so that the modified context information is changed into" as early as 2023, and the number of single young is reached.
Under the condition that the sensitive words are not related to the semantic information of the context information, the deletion of the sensitive words is not caused by incomplete semantics of the whole sentence, so that the sensitive words can be directly deleted, and news to be distributed after the sensitive words are deleted can be distributed.
In this embodiment, in the case that the news to be distributed includes the sensitive word, based on whether the sensitive word is related to the semantic information of the context information including the sensitive word, the processing mode of the sensitive word is determined, so that the use of the sensitive word is avoided while the integrity of the semantic of the context information is ensured as much as possible, and the experience of the user for reading the news is improved.
The news publishing device provided by the invention is described below, and the news publishing device described below and the news publishing method described above can be correspondingly referred to each other.
Fig. 5 is a schematic structural diagram of a news publishing device provided in an embodiment of the present invention, and as shown in fig. 5, the news publishing device 500 includes a first obtaining unit 501, a detecting unit 502, and a first publishing unit 503; wherein:
a first obtaining unit 501, configured to obtain news to be distributed;
the detection unit 502 is configured to input the news to be distributed into a news detection model, obtain a detection result output by the news detection model, where the detection result is used to represent the true and false situation of the news to be distributed, and the news detection model is obtained by training an initial news detection model based on sample news and tag information used to represent the true and false situation of the sample news;
a first publishing unit 503, configured to publish the news to be published if the detection result indicates that the news to be published is real news.
According to the news publishing device, when the news to be published is acquired, the news to be published is input into the news detection model, the detection result output by the news detection model is obtained, and when the detection result represents that the news to be published is real news, the news to be published is published. According to the news publishing method and device, when the news to be published is acquired, the true and false detection is carried out on the news to be published based on the news detection model, and the news to be published is published only under the condition that the news to be published is the true news, so that the accuracy of news publishing is improved.
Based on any of the above embodiments, the news bulletin device 500 further includes:
the second publishing unit is used for prohibiting the publishing of the news to be published and acquiring a target user identifier for inputting the news to be published under the condition that the detection result represents that the news to be published is false news;
The second acquisition unit is used for acquiring the quantity of false news input by the user represented by the target user identifier in a preset history time length;
And the adding unit is used for adding the target user identification into a blacklist under the condition that the number is larger than a first preset number.
Based on any one of the above embodiments, the second obtaining unit is specifically configured to:
determining the news type of the news to be distributed;
Acquiring the number of false news corresponding to the news type input by the user in a preset history time;
and adding the target user identifier to a blacklist under the condition that the number is larger than a first preset number, wherein the method comprises the following steps:
And adding the target user identifier into a blacklist under the condition that the number of false news corresponding to the news type is larger than the first preset number corresponding to the news type.
Based on any of the foregoing embodiments, the second obtaining unit is further specifically configured to:
Acquiring real reference news which has similar semantics to the news headlines of the news to be distributed and has similar semantics to news contents under the condition that the number is smaller than or equal to a first preset number;
And outputting prompt information, wherein the prompt information comprises real reference news, and the prompt information is used for reminding the user to modify the news to be distributed based on the real reference news.
Based on any of the above embodiments, the news bulletin device 500 further includes:
a third obtaining unit, configured to obtain comment information of the news to be distributed;
And the updating unit is used for updating the news detection model based on the news to be distributed and the label information representing that the news to be distributed is false news under the condition that the number of comment information representing that the news to be distributed is more than a second preset number.
Based on any of the above embodiments, the first distributing unit 503 is specifically configured to:
Determining whether the news to be distributed accords with a distribution rule under the condition that the detection result represents that the news to be distributed is real news;
determining whether the news to be distributed contains sensitive words or not based on a sensitive word stock under the condition that the news to be distributed accords with the distribution rule;
and publishing the news to be published under the condition that the news to be published does not contain sensitive words.
Based on any of the above embodiments, the first distributing unit 503 is further specifically configured to:
Acquiring context information containing sensitive words under the condition that the news to be distributed contains the sensitive words;
Determining whether the sensitive word is related to semantic information of the context information;
Under the condition that the sensitive word is related to semantic information of the context information, determining a target replacement word corresponding to the sensitive word based on a word replacement library, replacing the sensitive word with the target replacement word to obtain replaced news to be distributed, and distributing the replaced news to be distributed;
And deleting the sensitive words under the condition that the sensitive words are irrelevant to the semantic information of the context information, obtaining deleted news to be distributed, and distributing the deleted news to be distributed.
Fig. 6 is a schematic physical structure of an electronic device according to an embodiment of the present invention, as shown in fig. 6, the electronic device may include: processor 610, communication interface (Communications Interface) 620, memory 630, and communication bus 640, wherein processor 610, communication interface 620, memory 630 communicate with each other via communication bus 640. Processor 610 may invoke logic instructions in memory 630 to perform a news publishing method comprising: acquiring news to be distributed;
Inputting the news to be distributed into a news detection model to obtain a detection result output by the news detection model, wherein the detection result is used for representing the true and false conditions of the news to be distributed, and the news detection model is obtained by training an initial news detection model based on sample news and label information used for representing the true and false conditions of the sample news;
And publishing the news to be published under the condition that the detection result represents that the news to be published is real news.
Further, the logic instructions in the memory 630 may be implemented in the form of software functional units and stored in a computer-readable storage medium when sold or used as a stand-alone product. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a usb disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
In another aspect, the present invention also provides a computer program product, the computer program product comprising a computer program, the computer program being storable on a non-transitory computer readable storage medium, the computer program, when executed by a processor, being capable of executing the news distribution method provided by the above methods, the method comprising: acquiring news to be distributed;
Inputting the news to be distributed into a news detection model to obtain a detection result output by the news detection model, wherein the detection result is used for representing the true and false conditions of the news to be distributed, and the news detection model is obtained by training an initial news detection model based on sample news and label information used for representing the true and false conditions of the sample news;
And publishing the news to be published under the condition that the detection result represents that the news to be published is real news.
In yet another aspect, the present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, is implemented to perform the news distribution method provided by the above methods, the method comprising: acquiring news to be distributed;
Inputting the news to be distributed into a news detection model to obtain a detection result output by the news detection model, wherein the detection result is used for representing the true and false conditions of the news to be distributed, and the news detection model is obtained by training an initial news detection model based on sample news and label information used for representing the true and false conditions of the sample news;
And publishing the news to be published under the condition that the detection result represents that the news to be published is real news.
The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (10)
1. A news publishing method, comprising:
Acquiring news to be distributed;
Inputting the news to be distributed into a news detection model to obtain a detection result output by the news detection model, wherein the detection result is used for representing the true and false conditions of the news to be distributed, and the news detection model is obtained by training an initial news detection model based on sample news and label information used for representing the true and false conditions of the sample news;
And publishing the news to be published under the condition that the detection result represents that the news to be published is real news.
2. The news publishing method of claim 1, further comprising:
Under the condition that the detection result represents that the news to be distributed is false news, prohibiting the news to be distributed from being distributed, and acquiring a target user identification for inputting the news to be distributed;
acquiring the quantity of false news input by a user represented by the target user identifier in a preset history time;
And adding the target user identification into a blacklist under the condition that the number is larger than a first preset number.
3. The news publishing method of claim 2, wherein the obtaining the amount of false news entered by the user characterized by the target user identification for a preset historical time period comprises:
determining the news type of the news to be distributed;
Acquiring the number of false news corresponding to the news type input by the user in a preset history time;
and adding the target user identifier to a blacklist under the condition that the number is larger than a first preset number, wherein the method comprises the following steps:
And adding the target user identifier into a blacklist under the condition that the number of false news corresponding to the news type is larger than the first preset number corresponding to the news type.
4. The news publishing method of claim 2, further comprising:
Acquiring real reference news which has similar semantics to the news headlines of the news to be distributed and has similar semantics to news contents under the condition that the number is smaller than or equal to a first preset number;
And outputting prompt information, wherein the prompt information comprises real reference news, and the prompt information is used for reminding the user to modify the news to be distributed based on the real reference news.
5. The news distribution method according to any one of claims 1 to 4, characterized in that after the news to be distributed is distributed, the method further comprises:
Acquiring comment information of the news to be distributed;
And under the condition that the number of comment information representing that the news to be distributed is false news is larger than a second preset number, updating the news detection model based on the news to be distributed and label information representing that the news to be distributed is false news.
6. The news distribution method according to any one of claims 1-4, wherein, in the case where the detection result indicates that the news to be distributed is real news, distributing the news to be distributed includes:
Determining whether the news to be distributed accords with a distribution rule under the condition that the detection result represents that the news to be distributed is real news;
determining whether the news to be distributed contains sensitive words or not based on a sensitive word stock under the condition that the news to be distributed accords with the distribution rule;
and publishing the news to be published under the condition that the news to be published does not contain sensitive words.
7. The news publishing method of claim 6, further comprising:
Acquiring context information containing sensitive words under the condition that the news to be distributed contains the sensitive words;
Determining whether the sensitive word is related to semantic information of the context information;
Under the condition that the sensitive word is related to semantic information of the context information, determining a target replacement word corresponding to the sensitive word based on a word replacement library, replacing the sensitive word with the target replacement word to obtain replaced news to be distributed, and distributing the replaced news to be distributed;
And deleting the sensitive words under the condition that the sensitive words are irrelevant to the semantic information of the context information, obtaining deleted news to be distributed, and distributing the deleted news to be distributed.
8. A news bulletin device, comprising:
the first acquisition unit is used for acquiring news to be distributed;
the detection unit is used for inputting the news to be distributed into a news detection model to obtain a detection result output by the news detection model, wherein the detection result is used for representing the true and false conditions of the news to be distributed, and the news detection model is obtained by training an initial news detection model based on sample news and label information for representing the true and false conditions of the sample news;
the first issuing unit is used for issuing the news to be issued under the condition that the detection result represents that the news to be issued is real news.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the news distribution method of any one of claims 1 to 7 when the program is executed by the processor.
10. A non-transitory computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when executed by a processor, implements the news distribution method according to any one of claims 1 to 7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311837593.2A CN118035808A (en) | 2023-12-28 | 2023-12-28 | News release method, news release device, electronic equipment and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202311837593.2A CN118035808A (en) | 2023-12-28 | 2023-12-28 | News release method, news release device, electronic equipment and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN118035808A true CN118035808A (en) | 2024-05-14 |
Family
ID=90986750
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202311837593.2A Pending CN118035808A (en) | 2023-12-28 | 2023-12-28 | News release method, news release device, electronic equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN118035808A (en) |
-
2023
- 2023-12-28 CN CN202311837593.2A patent/CN118035808A/en active Pending
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20180349355A1 (en) | Artificial Intelligence Based Method and Apparatus for Constructing Comment Graph | |
US10423649B2 (en) | Natural question generation from query data using natural language processing system | |
US20090319449A1 (en) | Providing context for web articles | |
CN111079029B (en) | Sensitive account detection method, storage medium and computer equipment | |
CN112188312B (en) | Method and device for determining video material of news | |
CN113821605B (en) | Event extraction method | |
CN109800354B (en) | Resume modification intention identification method and system based on block chain storage | |
CN111782793A (en) | Intelligent customer service processing method, system and equipment | |
CN115238688B (en) | Method, device, equipment and storage medium for analyzing association relation of electronic information data | |
Jayatilaka et al. | Knowledge extraction for semantic web using web mining | |
CN117351336A (en) | Image auditing method and related equipment | |
CN113934834A (en) | Question matching method, device, equipment and storage medium | |
CN116662555B (en) | Request text processing method and device, electronic equipment and storage medium | |
CN113609390A (en) | Information analysis method and device, electronic equipment and computer readable storage medium | |
CN115248855A (en) | Text processing method and device, electronic equipment and computer readable storage medium | |
CN117113973A (en) | Information processing method and related device | |
CN110941713A (en) | Self-optimization financial information plate classification method based on topic model | |
CN118035808A (en) | News release method, news release device, electronic equipment and storage medium | |
CN113515949A (en) | Weakly supervised semantic entity recognition using general and target domain knowledge | |
CN113704422A (en) | Text recommendation method and device, computer equipment and storage medium | |
CN113505889B (en) | Processing method and device of mapping knowledge base, computer equipment and storage medium | |
CN111767727B (en) | Data processing method and device | |
CN113505889A (en) | Processing method and device of atlas knowledge base, computer equipment and storage medium | |
CN118133948A (en) | Question and answer method, device, equipment and storage medium based on large language model | |
CN118277518A (en) | Question answering method, device, equipment and medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |