CN106570162B - Artificial intelligence-based rumor recognition method and device - Google Patents
Artificial intelligence-based rumor recognition method and device Download PDFInfo
- Publication number
- CN106570162B CN106570162B CN201610974822.9A CN201610974822A CN106570162B CN 106570162 B CN106570162 B CN 106570162B CN 201610974822 A CN201610974822 A CN 201610974822A CN 106570162 B CN106570162 B CN 106570162B
- Authority
- CN
- China
- Prior art keywords
- text
- recognized
- rumor
- word
- vector
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 63
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 53
- 239000013598 vector Substances 0.000 claims abstract description 174
- 239000011159 matrix material Substances 0.000 claims abstract description 65
- 238000012549 training Methods 0.000 claims description 27
- 230000011218 segmentation Effects 0.000 claims description 21
- 238000012545 processing Methods 0.000 claims description 9
- 238000006243 chemical reaction Methods 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 4
- 230000008859 change Effects 0.000 claims description 4
- 230000006870 function Effects 0.000 description 51
- 238000010586 diagram Methods 0.000 description 9
- 238000003860 storage Methods 0.000 description 8
- 238000004590 computer program Methods 0.000 description 7
- 238000004422 calculation algorithm Methods 0.000 description 6
- 238000005457 optimization Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 230000003044 adaptive effect Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000002411 adverse Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/951—Indexing; Web crawling techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/284—Lexical analysis, e.g. tokenisation or collocates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a rumor identification method and a rumor identification device based on artificial intelligence, wherein the method comprises the following steps: acquiring a text to be identified; generating a word vector corresponding to the text to be recognized based on the bag of words BOW model; converting the word vector into a vector with the length of 2 by using a projection matrix model; taking a vector with the length of 2 as an input, calculating the probability that the text to be recognized is a rumor through a regression function SOFTMAX. By the method and the device, the network rumors in the internet information can be quickly identified, and the identification rate and timeliness of the network rumors are improved.
Description
Technical Field
The invention relates to the technical field of internet, in particular to a rumor identification method and device based on artificial intelligence.
Background
Artificial Intelligence (Artificial Intelligence), abbreviated in english as AI. The method is a new technical science for researching and developing theories, methods, technologies and application systems for simulating, extending and expanding human intelligence. Artificial intelligence is a branch of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can react in a manner similar to human intelligence, a field of research that includes robotics, language recognition, image recognition, natural language processing, and expert systems, among others. The development of artificial intelligence technology has also led to the progress of other related technologies, such as network rumor identification technology.
Network rumors refer to unpractical utterances propagated through network media (e.g., electronic mailboxes, chat software, social networking sites, web forums, etc.), primarily related to emergencies, public areas, celebrities, subversive traitors, etc. The propagation of network rumors is easy to disturb normal social order, and has adverse effect on society.
With the continuous development of internet technology, the transmission speed of internet information is faster and faster, and the transmission speed of network rumors is also faster. Therefore, how to effectively identify the network rumors in the internet information becomes a problem to be solved urgently in the technical field of the internet.
In the conventional network rumor identification method, whether the network information is a rumor is usually determined according to a preset keyword list. When the internet information has words matched with the words in the keyword list, the internet information is considered as a rumor. Because the conventional network rumor identification method is used for identifying the network rumors by presetting keywords, the identification rate is low, and the timeliness of the conventional network rumor identification method for identifying the rumors is poor due to the hysteresis of the keyword list.
Disclosure of Invention
The object of the present invention is to solve at least to some extent one of the above mentioned technical problems.
Therefore, a first objective of the present invention is to provide a rumor identification method based on artificial intelligence, which can quickly identify a network rumor in internet information, and improve the identification rate and timeliness of the network rumor.
The second objective of the present invention is to provide a rumor recognition device based on artificial intelligence.
A third object of the present invention is to provide a terminal.
A fourth object of the invention is to propose a non-transitory computer-readable storage medium.
A fifth object of the invention is to propose a computer program product.
In order to achieve the above object, an embodiment of a first aspect of the present invention provides a rumor identification method based on artificial intelligence, including: acquiring a text to be identified; generating a word vector corresponding to the text to be recognized based on the bag of words BOW model; converting the word vector into a vector with the length of 2 by using a projection matrix model; taking a vector with the length of 2 as an input, calculating the probability that the text to be recognized is a rumor through a regression function SOFTMAX.
In the artificial intelligence-based rumor recognition method provided by the embodiment of the first aspect of the invention, the text to be recognized is obtained, the word vector corresponding to the text to be recognized is generated based on the bag-of-words model, the word vector is converted into the vector with the length of 2 by using the projection matrix model, and the probability that the text to be recognized is the rumor is calculated through the regression function as input. Therefore, the network rumors in the internet information can be quickly identified, and the identification rate and the timeliness of the network rumors are improved.
In order to achieve the above object, a second embodiment of the present invention provides a rumor recognition apparatus based on artificial intelligence, including: the acquisition module is used for acquiring a text to be recognized; the generating module is used for generating word vectors corresponding to the texts to be recognized based on the bag of words BOW model; the conversion module is used for converting the word vector into a vector with the length of 2 by utilizing a projection matrix model; and the calculating module is used for taking the vector with the length of 2 as input and calculating the probability that the text to be recognized is the rumor through a regression function SOFTMAX.
In the artificial intelligence-based rumor recognition apparatus according to the second aspect of the present invention, the text to be recognized is obtained, the word vector corresponding to the text to be recognized is generated based on the bag-of-words model, the word vector is converted into a vector with a length of 2 by using the projection matrix model, and the probability that the text to be recognized is a rumor is calculated through the regression function as an input. Therefore, the network rumors in the internet information can be quickly identified, and the identification rate and the timeliness of the network rumors are improved.
In order to achieve the above object, an embodiment of a third aspect of the present invention provides a terminal, including: a processor; a memory for storing processor-executable instructions. Wherein the processor is configured to perform the steps of:
acquiring a text to be identified;
generating a word vector corresponding to the text to be recognized based on the bag of words BOW model;
converting the word vector into a vector with the length of 2 by using a projection matrix model;
taking a vector with the length of 2 as an input, calculating the probability that the text to be recognized is a rumor through a regression function SOFTMAX.
In the terminal provided by the embodiment of the third aspect of the present invention, the text to be recognized is obtained, the word vector corresponding to the text to be recognized is generated based on the bag-of-words model, the word vector is converted into the vector with the length of 2 by using the projection matrix model, and the probability that the text to be recognized is a rumor is calculated through the regression function as an input. Therefore, the network rumors in the internet information can be quickly identified, and the identification rate and the timeliness of the network rumors are improved.
In order to achieve the above object, a fourth aspect of the present invention provides a non-transitory computer-readable storage medium storing one or more programs which, when executed by a processor of a mobile terminal, enable the mobile terminal to perform an artificial intelligence based rumor recognition method, the method comprising;
acquiring a text to be identified;
generating a word vector corresponding to the text to be recognized based on the bag of words BOW model;
converting the word vector into a vector with the length of 2 by using a projection matrix model;
taking a vector with the length of 2 as an input, calculating the probability that the text to be recognized is a rumor through a regression function SOFTMAX.
In the non-transitory computer-readable storage medium according to the fourth aspect of the present invention, a text to be recognized is obtained, a word vector corresponding to the text to be recognized is generated based on a bag-of-words model, the word vector is converted into a vector with a length of 2 by using a projection matrix model, and the probability that the text to be recognized is a rumor is calculated through a regression function as an input. Therefore, the network rumors in the internet information can be quickly identified, and the identification rate and the timeliness of the network rumors are improved.
In order to achieve the above object, a fifth embodiment of the present invention provides a computer program product, wherein instructions of the computer program product, when executed by a processor, perform an artificial intelligence-based rumor recognition method, the method comprising:
acquiring a text to be identified;
generating a word vector corresponding to the text to be recognized based on the bag of words BOW model;
converting the word vector into a vector with the length of 2 by using a projection matrix model;
taking a vector with the length of 2 as an input, calculating the probability that the text to be recognized is a rumor through a regression function SOFTMAX.
In the computer program product according to the fifth embodiment of the present invention, a text to be recognized is obtained, a word vector corresponding to the text to be recognized is generated based on a bag-of-words model, the word vector is converted into a vector with a length of 2 by using a projection matrix model, and the probability that the text to be recognized is a rumor is calculated through a regression function as an input. Therefore, the network rumors in the internet information can be quickly identified, and the identification rate and the timeliness of the network rumors are improved.
Drawings
The foregoing and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
fig. 1 is a schematic flow chart illustrating an artificial intelligence-based rumor identification method according to an embodiment of the present invention;
FIG. 2 is a schematic flow chart of generating word vectors corresponding to a text to be recognized based on a BOW model;
fig. 3 is an exemplary diagram illustrating the embodiment by taking a text to be recognized as a web article content as an example;
fig. 4 is a schematic flow chart illustrating an artificial intelligence-based rumor identification method according to another embodiment of the present invention;
FIG. 5 is a schematic flow chart of training parameters of a projection matrix model;
fig. 6 is a schematic flow chart illustrating an artificial intelligence-based rumor identification method according to another embodiment of the present invention;
fig. 7 is a schematic structural diagram of a rumor recognition apparatus based on artificial intelligence according to an embodiment of the present invention;
fig. 8 is a schematic structural diagram of a rumor recognition apparatus based on artificial intelligence according to another embodiment of the present invention;
fig. 9 is a schematic structural diagram of a rumor recognition apparatus based on artificial intelligence according to another embodiment of the present invention;
fig. 10 is a schematic structural diagram of a rumor recognition apparatus based on artificial intelligence according to still another embodiment of the present invention;
fig. 11 is a schematic structural diagram of a rumor recognition apparatus based on artificial intelligence according to still another embodiment of the present invention.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative and intended to be illustrative of the invention and are not to be construed as limiting the invention. On the contrary, the embodiments of the invention include all changes, modifications and equivalents coming within the spirit and terms of the claims appended hereto.
Fig. 1 is a schematic flow chart of a rumor identification method based on artificial intelligence according to an embodiment of the present invention.
As shown in fig. 1, the artificial intelligence-based rumor identification method of the present embodiment includes:
s11: and acquiring a text to be recognized.
In this embodiment, in order to determine whether a piece of internet information is a rumor, the piece of internet information is first acquired through the internet as a text to be recognized.
The internet information may be long-spread web articles, or news headlines.
S12: and generating a word vector corresponding to the text to be recognized based on the bag of words BOW model.
In this embodiment, after the text to be recognized is obtained, a word vector corresponding to the text to be recognized may be generated according to a Bag Of Words (BOW) model.
The BOW model is a document representation method commonly used in the field of information retrieval. In information retrieval, the BOW model assumes that a document is only regarded as a collection of vocabularies without considering the elements of word sequence, grammar, syntax and the like of the document, and the appearance of each word in the document is independent and independent of whether other words appear or not.
Let us assume that several documents are contained in a large collection of documents. All words in all documents are extracted to form a dictionary containing Q words. Using the BOW model, each document can be represented as a vector in Q dimensions, Q being a positive integer. Wherein the ith element in the vector represents the number of times the ith word in the dictionary appears in the document, and i is a positive integer.
Specifically, as shown in fig. 2, generating a word vector corresponding to a text to be recognized based on a BOW model may include the following steps:
s121: and segmenting the text to be recognized into a plurality of word segments.
In this embodiment, in order to obtain a word vector corresponding to a text to be recognized based on a BOW model, after the text to be recognized is obtained, a word segmentation process is performed on the text to be recognized by using a correlation technique, and the text to be recognized is segmented into a plurality of words.
S122: and acquiring word segmentation vectors corresponding to the plurality of word segmentations.
In this embodiment, after the text to be recognized is subjected to word segmentation processing, word segmentation vectors corresponding to the respective word segmentations are further obtained. The word segmentation vector corresponding to each word segmentation can be obtained by searching a dictionary.
In particular, assuming that there is a dictionary containing N words, each of which has a word vector size (embeddingsize) of M, the dictionary can be represented as a word vector matrix of N × M. Wherein N, M is a positive integer, and the size of M is usually set to be 50-1000. For a given word, the word vector for the word is obtained by looking up the word vector matrix, assuming that the word corresponds to the kth row in the word vector matrix. Wherein k is a positive integer.
Take the text to be recognized as the content of the web article as an example. Firstly, content is divided into a plurality of participles by adopting the correlation technique, and the participles are respectively marked as w1，w2，…，wn. Then, according to the word vector matrix, the word vectors corresponding to the n participles can be obtained by looking up the dictionary, and are respectively marked as emb (w)1)，emb(w2)，…，emb(wn)。
It should be noted that a word vector is a way to digitize words in a language, i.e., to represent a word as a vector. The simplest word vector Representation method is One-hot Representation, which represents each word as a very long vector, the dimension of the vector represents the size of a word list, the vector component value of only One dimension is '1', the other component values are all '0', and the position of '1' corresponds to the position of the word in the word list. For example, "microphone" is denoted as [ 0000000010000000. ], and assuming that it is recorded from 0, the microphone is denoted as 8, meaning that the word microphone is at the 8 th position in the vocabulary. Another word vector Representation method is Distributed Representation, and its basic idea is: by training each word in a language to map to a fixed-length vector, the set of all vectors forms a word vector space, where each vector represents a point in the space. The concept of "distance" is introduced into the word vector space, that is, the similarity between words in syntax and semantics can be judged according to the distance between words. The present invention may represent the word vector in any way, and is not limited thereto.
S123: and operating the word vector by using the BOW model to generate a word vector corresponding to the text to be recognized.
In this embodiment, after the word segmentation vector is obtained, the word segmentation vector may be operated by using a BOW model to generate a word vector corresponding to the text to be recognized.
Specifically, the BOW model performs simple summation operation on the word segmentation vectors corresponding to the word segmentation, that is, performs element-by-element addition on the word segmentation vectors, and the obtained result is the word vector corresponding to the text to be recognized.
Therefore, the word vector rep (content) corresponding to the content can be expressed as:
rep(content)＝sum(emb(w1)，emb(w2)，…，emb(wn))
s13: the word vectors are converted to length 2 vectors using a projection matrix model.
In this embodiment, after the word vector corresponding to the text to be recognized is obtained, the word vector may be converted into a vector with a length of 2 by using the projection matrix model.
Specifically, converting the word vector into a length-2 vector using a projection matrix model includes: performing projection operation on the word vectors by using a projection matrix model to generate a matrix corresponding to the word vectors; the matrix is operated by a nonlinear change function to generate a vector of length 2. Wherein the nonlinear variation function comprises one of a sigmoid function, a tangent function and an activation function.
It should be noted that the operation of projection operation and operation on the matrix by the nonlinear variation function is not limited to one time, and a vector with a length of 2 may be obtained by multiple operations, which is not limited in the present invention.
Still taking the text to be recognized as the content of the web article as an example, the word vector rep (content) corresponding to the content is obtained by using the BOW model. rep (content) is a vector with length of N, and the projection matrix model is used to perform projection operation on rep (content), that is, the matrix of N × M is multiplied by rep (content) to obtain a vector with length of M. Then, the obtained vector is subjected to nonlinear operation through a nonlinear variation function such as a sigmoid function, and the obtained vector is still a vector with the length of M. And continuously performing projection operation on the vector obtained after the nonlinear operation, and multiplying the vector by a matrix of M x 2 to obtain a vector with the length of 2. Wherein M, N is a positive integer.
S14: taking a vector with the length of 2 as an input, calculating the probability that the text to be recognized is a rumor through a regression function SOFTMAX.
In this embodiment, after the word vector corresponding to the text to be recognized is converted into the vector with the length of 2, the obtained vector is used as an input, and the probability that the text to be recognized is a rumor can be calculated and obtained through the regression function SOFTMAX.
The SOFTMAX function is a function capable of solving a multi-classification problem, and essentially maps any real number vector of one K dimension into a real number vector of another K dimension, wherein each element in the vector has a value between (0 and 1), the sum of all elements in the vector is 1, and K is a positive integer. Therefore, when a vector with the length of 2 is input, the vector with the length of 2 is also output through the SOFTMAX function, values of two elements in the vector are both 0-1, and the sum of the two vectors is 1, so that the probability that the text to be recognized is a rumor and the probability that the text is not a rumor can be respectively represented.
For example, for the SOFTMAX function, its outputOut can be expressed as y ═ y1，y2，…，yk]Where k is a positive integer representing the length of the output vector. In this example, k is 2. Suppose that the first element of the length-2 vector output by the SOFTMAX function represents the probability that the text to be recognized is not a rumor, and the second element represents the probability that the text to be recognized is a rumor. If the output vector of the SOFTMAX function is y ═ 0.8, 0.2 for a certain text to be recognized]That is, the probability of not being a rumor is 0.8, and the probability of being a rumor is 0.2, it indicates that the text to be recognized is not a rumor. If the output vector of the SOFTMAX function is y ═ 0.26, 0.74 for a certain text to be recognized]If the probability of not being a rumor is 0.26 and the probability of being a rumor is 0.74, the text to be recognized is a rumor.
The artificial intelligence-based rumor identification method of the embodiment may be deployed in a content production server (e.g., a post), a content forwarding server (e.g., a compulsory binding), and other servers, and used to determine whether internet information is a rumor.
The following describes the embodiment specifically by taking the text to be recognized as the content of the web article. As shown in fig. 3, in this embodiment, an existing official report sample and a rumor sample are first obtained from an internet database, and are respectively labeled as a positive example and a negative example as training samples, and are trained by using a gradient-based model to obtain parameters of a projection matrix model. After obtaining a certain network article content in the internet, generating word vector representation of the network article content based on a BOW model, further combining the parameters of a projection matrix model obtained by training, converting the word vector representation of the network article content into a vector with the length of 2, using the vector as the input of an SOFTMAX function, and finally judging whether the network article content is a rumor according to the output of the SOFTMAX function. The artificial intelligence-based rumor recognition method provided by the embodiment of the invention generates word vectors corresponding to the text to be recognized based on the bag-of-words model by acquiring the text to be recognized, converts the word vectors into vectors with the length of 2 by using the projection matrix model, and calculates the probability that the text to be recognized is a rumor through a regression function as input. Therefore, the network rumors in the internet information can be quickly identified, and the identification rate and the timeliness of the network rumors are improved.
Fig. 4 is a schematic flow chart illustrating an artificial intelligence-based rumor identification method according to another embodiment of the present invention.
As shown in fig. 4, based on the above embodiment, before converting the word vector into a length-2 vector by using the projection matrix model, the following steps may be further included:
s15: and training parameters of the projection matrix model.
In this embodiment, in order to convert the word vector of the text to be recognized into a vector with a length of 2 by using the projection matrix model, the parameters of the projection matrix model need to be trained first.
It should be noted that, the parameters for training the projection matrix model are not necessarily performed after step S12, and may be performed at any time before step S13 is performed, which is not limited by the invention.
Specifically, as shown in fig. 5, training the parameters of the projection matrix model may include the following steps:
s151: sample data was obtained, including official report samples and rumor samples.
In this embodiment, sample data may be acquired from an internet database as training data required for training parameters of a projection matrix model, where the sample data includes official report samples and rumor samples.
S152: the official report samples were taken as positive examples and rumor samples as negative examples, the generated parameters were trained, and the parameters were optimized using a gradient-based model.
In this embodiment, after sample data is obtained, the obtained official report sample is labeled as a positive example, the obtained rumor sample is labeled as a negative example, the labeled sample data is used as training data after the labeling is completed, parameters of a projection matrix model are generated through training, and then the generated parameters are optimized by using a gradient-based model.
Specifically, sample data may be trained to generate parameters of the projection matrix model by using a pair-wise pair-based pair-wise training method or by using a point-wise training method based on a single sample. There may also be various methods for optimizing parameters by using a Gradient-based model, such as a Stochastic Gradient Descent (SGD) algorithm, a moment (Momentum) algorithm, an Adaptive Gradient (Adaptive Gradient) algorithm, a Back Propagation (BP) algorithm, and the like.
Take SGD optimization algorithm as an example. The idea of the SGD algorithm is to iteratively update the parameters of the generated projection matrix model by computing the gradients (partial derivatives of the parameters) of a certain set of sample data. The process of iterative update is: the gradient of the parameter obtained in the previous iteration is multiplied by a learning rate (learning rate), i.e., a step size, and the result obtained in the current iteration is updated to the parameter. After a plurality of iterations, the difference between the value of the finally obtained parameter and the actual value can be converged to a negative log loss function.
It should be noted that parameters of the projection matrix model may be generated by training sample data by using a pair-wise method or a point-wise method, or may be generated by using other training methods. In addition, other loss functions may be used as optimization objectives, such as 0-1 loss functions, squared loss functions, absolute loss functions, and the like. The invention does not limit the training method, the optimization method and the optimization objective function of the parameters.
According to the artificial intelligence-based rumor recognition method provided by the embodiment of the invention, the official report sample and the rumor sample are obtained and used as sample data and used as parameters for training the projection matrix model respectively in the positive case and the negative case, and the model optimization parameters based on the gradient can enable the operation result of the projection matrix model to be more accurate, and further improve the accuracy of network rumor recognition.
Fig. 6 is a flowchart illustrating a rumor identification method based on artificial intelligence according to another embodiment of the present invention.
As shown in fig. 6, based on the above embodiment, after calculating the probability that the text to be recognized is a rumor, the method may further include the following steps:
s16: and correspondingly processing the text to be recognized according to the probability.
In this embodiment, after the probability that the text to be recognized is a rumor is calculated, whether the text to be recognized is a rumor can be determined, and the text to be recognized is correspondingly processed according to the determination result. For example, if the text to be recognized is determined to be a rumor, the text to be recognized may be subjected to processing such as account number prohibition and eye-catching mark; and if the text to be recognized is determined not to be a rumor, directly displaying the text to be recognized.
According to the artificial intelligence-based rumor recognition method provided by the embodiment of the invention, after the probability that the text to be recognized is the rumor is calculated, the text to be recognized is correspondingly processed according to the obtained probability, so that the user can be helped to recognize the network rumor, and the user is prevented from being damaged by rumor information.
In order to implement the foregoing embodiments, the present invention further provides a rumor recognition apparatus based on artificial intelligence, and fig. 7 is a schematic structural diagram of the rumor recognition apparatus based on artificial intelligence according to an embodiment of the present invention.
As shown in fig. 7, the artificial intelligence-based rumor recognition apparatus of the present embodiment includes: an acquisition module 710, a generation module 720, a conversion module 730, and a calculation module 740. Wherein,
the obtaining module 710 is configured to obtain a text to be recognized.
And a generating module 720, configured to generate a word vector corresponding to the text to be recognized based on the bag of words BOW model.
Specifically, as shown in fig. 8, the generating module 720 includes:
the segmentation unit 721 is configured to segment the text to be recognized into a plurality of segments.
The first obtaining unit 722 is configured to obtain a participle vector corresponding to a plurality of participles.
And the operation unit 723 is configured to perform operation on the word vector by using the BOW model to generate a word vector corresponding to the text to be recognized.
A converting module 730, configured to convert the word vector into a length-2 vector by using a projection matrix model.
Specifically, the conversion module 730 is configured to:
performing projection operation on the word vectors by using a projection matrix model to generate a matrix corresponding to the word vectors;
the matrix is operated by a nonlinear change function to generate a vector of length 2.
Wherein the nonlinear variation function comprises one of a sigmoid function, a tangent function and an activation function.
The calculating module 740 is configured to use the length-2 vector as an input, and calculate the probability that the text to be recognized is a rumor through a regression function SOFTMAX.
It should be noted that the explanation of the artificial intelligence based rumor identification method in the foregoing embodiments is also applicable to the artificial intelligence based rumor identification apparatus in this embodiment, and the implementation principle is similar, and is not repeated here.
The artificial intelligence-based rumor recognition device provided by the embodiment of the invention generates word vectors corresponding to the text to be recognized based on the bag-of-words model by acquiring the text to be recognized, converts the word vectors into vectors with the length of 2 by using the projection matrix model, and calculates the probability that the text to be recognized is a rumor through a regression function as input. Therefore, the network rumors in the internet information can be quickly identified, and the identification rate and the timeliness of the network rumors are improved.
Fig. 9 is a schematic structural diagram of a rumor recognition apparatus based on artificial intelligence according to another embodiment of the present invention. As shown in fig. 9, based on the description in fig. 7, the artificial intelligence-based rumor recognition apparatus according to the present embodiment may further include:
and a training module 750 for training parameters of the projection matrix model.
Specifically, as shown in fig. 10, the training module 750 includes:
a second obtaining unit 751 for obtaining sample data, the sample data comprising official report samples and rumor samples.
A training unit 752 for training the generated parameters with the official report samples as positive examples and rumor samples as negative examples, and optimizing the parameters using the gradient-based model.
It should be noted that the explanation of the artificial intelligence based rumor identification method in the foregoing embodiments is also applicable to the artificial intelligence based rumor identification apparatus in this embodiment, and the implementation principle is similar, and is not repeated here.
According to the artificial intelligence-based rumor recognition device provided by the embodiment of the invention, the official report sample and the rumor sample are obtained and used as sample data and used as parameters for training the projection matrix model respectively in the positive case and the negative case, and the model optimization parameters based on the gradient can enable the operation result of the projection matrix model to be more accurate, and further improve the accuracy of network rumor recognition.
Fig. 11 is a schematic structural diagram of a rumor recognition apparatus based on artificial intelligence according to still another embodiment of the present invention. As shown in fig. 11, based on the description in fig. 7, the artificial intelligence-based rumor recognition apparatus according to the present embodiment may further include:
the processing module 760 is configured to, after calculating the probability that the text to be recognized is a rumor, perform corresponding processing on the text to be recognized according to the probability.
In this embodiment, after the probability that the text to be recognized is a rumor is calculated, whether the text to be recognized is a rumor can be determined, and the text to be recognized is correspondingly processed according to the determination result. For example, if the text to be recognized is determined to be a rumor, the text to be recognized may be subjected to processing such as account number prohibition and eye-catching mark; and if the text to be recognized is determined not to be a rumor, directly displaying the text to be recognized.
It should be noted that the explanation of the artificial intelligence based rumor identification method in the foregoing embodiments is also applicable to the artificial intelligence based rumor identification apparatus in this embodiment, and the implementation principle is similar, and is not repeated here.
According to the artificial intelligence-based rumor recognition device provided by the embodiment of the invention, after the probability that the text to be recognized is the rumor is calculated, the text to be recognized is correspondingly processed according to the obtained probability, so that the user can be helped to recognize the network rumor, and the user is prevented from being damaged by rumor information.
In order to implement the above embodiments, the present invention further provides a terminal, including: a processor, and a memory for storing processor-executable instructions. Wherein the processor is configured to perform the steps of:
s11': and acquiring a text to be recognized.
S12': and generating a word vector corresponding to the text to be recognized based on the bag of words BOW model.
S13': the word vectors are converted to length 2 vectors using a projection matrix model.
S14': taking a vector with the length of 2 as an input, calculating the probability that the text to be recognized is a rumor through a regression function SOFTMAX.
It should be noted that the explanation of the artificial intelligence-based rumor identification method in the foregoing embodiment is also applicable to the terminal in this embodiment, and the implementation principle is similar, and is not described herein again.
According to the terminal provided by the embodiment of the invention, the text to be recognized is obtained, the word vector corresponding to the text to be recognized is generated based on the bag-of-words model, the word vector is converted into the vector with the length of 2 by using the projection matrix model, and the probability that the text to be recognized is a rumor is calculated through a regression function as input. Therefore, the network rumors in the internet information can be quickly identified, and the identification rate and the timeliness of the network rumors are improved.
In order to implement the above embodiments, the present invention further provides a non-transitory computer-readable storage medium storing one or more programs, which when executed by a processor of a mobile terminal, enable the mobile terminal to perform the artificial intelligence based rumor recognition method according to the first aspect of the present invention.
The non-transitory computer-readable storage medium provided in the embodiment of the present invention obtains a text to be recognized, generates a word vector corresponding to the text to be recognized based on a bag-of-words model, converts the word vector into a vector with a length of 2 using a projection matrix model, and calculates a probability that the text to be recognized is a rumor through a regression function as an input. Therefore, the network rumors in the internet information can be quickly identified, and the identification rate and the timeliness of the network rumors are improved.
In order to implement the foregoing embodiments, the present invention further provides a computer program product, wherein when instructions of the computer program product are executed by a processor, the method for artificial intelligence based rumor recognition according to the first aspect of the present invention is performed.
The computer program product provided by the embodiment of the invention generates word vectors corresponding to the text to be recognized based on the bag-of-words model by acquiring the text to be recognized, converts the word vectors into vectors with the length of 2 by using the projection matrix model, and calculates the probability that the text to be recognized is a rumor through a regression function as input. Therefore, the network rumors in the internet information can be quickly identified, and the identification rate and the timeliness of the network rumors are improved.
It should be noted that the terms "first," "second," and the like in the description of the present invention are used for descriptive purposes only and are not to be construed as indicating or implying relative importance. In addition, in the description of the present invention, "a plurality" means two or more unless otherwise specified.
Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing specific logical functions or steps of the process, and alternate implementations are included within the scope of the preferred embodiment of the present invention in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present invention.
It should be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the various steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
It will be understood by those skilled in the art that all or part of the steps carried by the method for implementing the above embodiments may be implemented by hardware related to instructions of a program, which may be stored in a computer readable storage medium, and when the program is executed, the program includes one or a combination of the steps of the method embodiments.
In addition, functional units in the embodiments of the present invention may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a stand-alone product, may also be stored in a computer readable storage medium.
The storage medium mentioned above may be a read-only memory, a magnetic or optical disk, etc.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.
Claims (10)
1. A rumor identification method based on artificial intelligence is characterized by comprising the following steps:
acquiring a text to be identified;
generating a word vector corresponding to the text to be recognized based on a bag of words BOW model;
converting the word vector into a vector with the length of 2 by using a projection matrix model, wherein the parameters of the projection matrix model are obtained by training in the following way: obtaining sample data, wherein the sample data comprises an official report sample and a rumor sample; training to generate the parameters by taking the official report sample as a positive example and the rumor sample as a negative example, and optimizing the parameters by using a gradient-based model;
and taking the vector with the length of 2 as an input, and calculating the probability that the text to be recognized is a rumor through a regression function SOFTMAX.
2. The method of claim 1, wherein generating a word vector corresponding to the text to be recognized based on a bag of words model BOW comprises:
segmenting the text to be recognized into a plurality of word segments;
acquiring word segmentation vectors corresponding to the plurality of word segmentations;
and operating the word segmentation vectors by using the BOW model to generate word vectors corresponding to the text to be recognized.
3. The method of claim 1, wherein converting the word vector to a length-2 vector using a projection matrix model comprises:
performing projection operation on the word vectors by using a projection matrix model to generate a matrix corresponding to the word vectors;
and operating the matrix through a nonlinear change function to generate the vector with the length of 2.
4. The method of claim 3, wherein the non-linear variation function comprises one of a sigmoid function, a tangent function, and an activation function.
5. The method of claim 1, further comprising:
after the probability that the text to be recognized is a rumor is calculated, the text to be recognized is correspondingly processed according to the probability.
6. An artificial intelligence-based rumor recognition device, comprising:
the acquisition module is used for acquiring a text to be recognized;
the generating module is used for generating a word vector corresponding to the text to be recognized based on a bag of words BOW model;
a conversion module, configured to convert the word vector into a vector with a length of 2 by using a projection matrix model, where parameters of the projection matrix model are obtained through training in the following manner: obtaining sample data, wherein the sample data comprises an official report sample and a rumor sample; training to generate the parameters by taking the official report sample as a positive example and the rumor sample as a negative example, and optimizing the parameters by using a gradient-based model;
and the calculating module is used for taking the vector with the length of 2 as input and calculating the probability that the text to be recognized is a rumor through a regression function SOFTMAX.
7. The apparatus of claim 6, wherein the generating module comprises:
the segmentation unit is used for segmenting the text to be recognized into a plurality of word segments;
the first acquisition unit is used for acquiring word segmentation vectors corresponding to the plurality of word segmentations;
and the operation unit is used for operating the word segmentation vectors by utilizing the BOW model so as to generate word vectors corresponding to the text to be recognized.
8. The apparatus of claim 6, wherein the conversion module is to:
performing projection operation on the word vectors by using a projection matrix model to generate a matrix corresponding to the word vectors;
and operating the matrix through a nonlinear change function to generate the vector with the length of 2.
9. The apparatus of claim 8, in which the non-linear variation function comprises one of a sigmoid function, a tangent function, and an activation function.
10. The apparatus of claim 6, further comprising:
and the processing module is used for performing corresponding processing on the text to be recognized according to the probability after the probability that the text to be recognized is a rumor is calculated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610974822.9A CN106570162B (en) | 2016-11-04 | 2016-11-04 | Artificial intelligence-based rumor recognition method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610974822.9A CN106570162B (en) | 2016-11-04 | 2016-11-04 | Artificial intelligence-based rumor recognition method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106570162A CN106570162A (en) | 2017-04-19 |
| CN106570162B true CN106570162B (en) | 2020-07-28 |
Family
ID=58539915
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610974822.9A Active CN106570162B (en) | 2016-11-04 | 2016-11-04 | Artificial intelligence-based rumor recognition method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106570162B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109753646B (en) * | 2017-11-01 | 2022-10-21 | 深圳市腾讯计算机系统有限公司 | Article attribute identification method and electronic equipment |
| CN108491480B (en) * | 2018-03-12 | 2021-05-11 | 义语智能科技（上海）有限公司 | Rumor detection method and apparatus |
| CN108614855A (en) * | 2018-03-19 | 2018-10-02 | 众安信息技术服务有限公司 | A kind of rumour recognition methods |
| CN109388696B (en) * | 2018-09-30 | 2021-07-23 | 北京字节跳动网络技术有限公司 | Method, device, storage medium and electronic equipment for deleting rumor article |
| CN113535944A (en) * | 2020-04-21 | 2021-10-22 | 阿里巴巴集团控股有限公司 | Text processing method and device, electronic equipment and computer readable storage medium |
| CN111610913B (en) * | 2020-04-24 | 2021-08-24 | 维沃移动通信有限公司 | Message identification method and device and electronic equipment |
| CN112487176B (en) * | 2020-11-26 | 2021-11-02 | 北京智谱华章科技有限公司 | Social robot detection method, system, storage medium and electronic device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103902621A (en) * | 2012-12-28 | 2014-07-02 | 深圳先进技术研究院 | Method and device for identifying network rumor |
| CN104834747A (en) * | 2015-05-25 | 2015-08-12 | 中国科学院自动化研究所 | Short text classification method based on convolution neutral network |
| CN105045857A (en) * | 2015-07-09 | 2015-11-11 | 中国科学院计算技术研究所 | Social network rumor recognition method and system |
| CN105354305A (en) * | 2015-11-05 | 2016-02-24 | 北京邮电大学 | Online-rumor identification method and apparatus |
| CN105787101A (en) * | 2016-03-18 | 2016-07-20 | 联想(北京)有限公司 | Information processing method and electronic equipment |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104679739A (en) * | 2013-11-27 | 2015-06-03 | 江苏华御信息技术有限公司 | Method for controlling spreading of unreal information |
| JP6070951B2 (en) * | 2013-12-17 | 2017-02-01 | インターナショナル・ビジネス・マシーンズ・コーポレーションＩｎｔｅｒｎａｔｉｏｎａｌ Ｂｕｓｉｎｅｓｓ Ｍａｃｈｉｎｅｓ Ｃｏｒｐｏｒａｔｉｏｎ | Apparatus and method for supporting analysis of evaluation |
| US9984067B2 (en) * | 2014-04-18 | 2018-05-29 | Thomas A. Visel | Automated comprehension of natural language via constraint-based processing |
-
2016
- 2016-11-04 CN CN201610974822.9A patent/CN106570162B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103902621A (en) * | 2012-12-28 | 2014-07-02 | 深圳先进技术研究院 | Method and device for identifying network rumor |
| CN104834747A (en) * | 2015-05-25 | 2015-08-12 | 中国科学院自动化研究所 | Short text classification method based on convolution neutral network |
| CN105045857A (en) * | 2015-07-09 | 2015-11-11 | 中国科学院计算技术研究所 | Social network rumor recognition method and system |
| CN105354305A (en) * | 2015-11-05 | 2016-02-24 | 北京邮电大学 | Online-rumor identification method and apparatus |
| CN105787101A (en) * | 2016-03-18 | 2016-07-20 | 联想(北京)有限公司 | Information processing method and electronic equipment |
Non-Patent Citations (1)
| Title |
|---|
| "微博谣言识别研究";贺刚 等;《图书情报工作》;20131205;第57卷(第23期);第114-120页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106570162A (en) | 2017-04-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106570162B (en) | Artificial intelligence-based rumor recognition method and device | |
| CN107168952B (en) | Information generation method and device based on artificial intelligence | |
| CN106502985B (en) | neural network modeling method and device for generating titles | |
| CN106557563A (en) | Query statement based on artificial intelligence recommends method and device | |
| US20150095017A1 (en) | System and method for learning word embeddings using neural language models | |
| CN108733837A (en) | A kind of the natural language structural method and device of case history text | |
| KR20210124111A (en) | Method and apparatus for training model, device, medium and program product | |
| KR20190019661A (en) | Method for Natural Langage Understanding Based on Distribution of Task-specific Labels | |
| JP6291443B2 (en) | Connection relationship estimation apparatus, method, and program | |
| CN108664512B (en) | Text object classification method and device | |
| CN110263325A (en) | Chinese automatic word-cut | |
| CN112100401B (en) | Knowledge graph construction method, device, equipment and storage medium for science and technology services | |
| CN115244587A (en) | Efficient ground truth annotation | |
| WO2019160096A1 (en) | Relationship estimation model learning device, method, and program | |
| CN114995903B (en) | Class label identification method and device based on pre-training language model | |
| CN111611395A (en) | Entity relationship identification method and device | |
| CN111767694A (en) | Text generation method and device and computer readable storage medium | |
| CN116245097A (en) | Method for training entity recognition model, entity recognition method and corresponding device | |
| CN114722833A (en) | Semantic classification method and device | |
| CN116127060A (en) | Text classification method and system based on prompt words | |
| CN114444515A (en) | Relation extraction method based on entity semantic fusion | |
| JP2018041300A (en) | Machine learning model generation device and program | |
| CN112380861B (en) | Model training method and device and intention recognition method and device | |
| CN109117471B (en) | Word relevancy calculation method and terminal | |
| CN117034916A (en) | Method, device and equipment for constructing word vector representation model and word vector representation |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |