Cross-platform social robot detection method based on adversarial neural network
What is Al technical title?
Al technical title is built by PatSnap Al team. It summarizes the technical point description of the patent document.
A detection method and neural network technology, applied in the field of robot detection, to achieve good versatility
Pending Publication Date: 2021-06-25
TIANJIN UNIV
View PDF3 Cites 0 Cited by
Summary
Abstract
Description
Claims
Application Information
AI Technical Summary
This helps you quickly interpret patents by identifying the three key elements:
Problems solved by technology
Method used
Benefits of technology
Problems solved by technology
Although the current research on social robot detection has achieved certain results, these results are all realized around a social platform and can only identify specific types of social robots [2]
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more
Image
Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
Click on the blue label to locate the original text in one second.
Reading with bidirectional positioning of images and text.
Smart Image
Examples
Experimental program
Comparison scheme
Effect test
Embodiment 1
[0021] In order to solve the problem that all kinds of social robots emerge in an endless stream in social robot detection, the application scenarios of traditional algorithms are limited, see figure 1 The implementation of the present invention provides a method for detecting a cross-platform social robot based on an adversarial neural network, the method comprising the following steps:
[0022] 101: data preprocessing;
[0023] Among them, this step includes three sub-steps of data cleaning, data standardization, and attribute extraction, aiming to standardize the input format of the neural network and reduce abnormal conditions during the training process of the neural network.
[0024] 102: Account feature extraction;
[0025] First, 17 attribute values are extracted according to each account data, so as to abstractly represent an account information, and form a 1*17 one-dimensional vector, which is sent into a one-dimensional convolutional neural network, and in the vo...
Embodiment 2
[0029] The scheme in embodiment 1 is introduced in detail below in conjunction with specific calculation formulas and examples, see the following description for details:
[0030] 201: data cleaning;
[0031] Due to the data in the data set, there are some missing and incomplete attribute values and data with obvious abnormalities, which may lead to deviations in feature extraction. [8] , so it is necessary to fill in the missing values and at the same time identify outliers and discard them.
[0032] 202: data standardization;
[0033] Because the attribute values of data often have different dimensions, for example, the unit of the data attribute of the number of fans is 1, and the unit of the data attribute of the number of days of account creation is days, and the unit of the data attribute of whether to set the background of the personal homepage is Yes / No. In order to convert the data into dimensionless values, thereby getting rid of the limitation of the unit, s...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more
PUM
Login to view more
Abstract
The invention discloses a cross-platform social robot detection method based on an adversarial neural network, and the method comprises the steps: carrying out the processing of an input format of an account feature extractor through data cleaning, data standardization and attribute extraction; extracting a plurality of attribute values from each piece of account data to form a one-dimensional vector, transmitting the vector as input into an account feature extractor, adding a batch normalization processing layer behind a convolutional layer of the account feature extractor to accelerate the extraction process of account features, and extracting a hidden layer as depth feature representation of an account; for data from different social network platforms, training an account feature extractor and a domain discriminator at the same time in a confrontation mode to align cross-platform data feature distribution, and achieving cross-platform social robot detection. According to the invention, the universality of the social robot detection algorithm is improved.
Description
technical field [0001] The invention relates to the field of robot detection, in particular to a cross-platform social robot detection method based on an adversarial neural network. Background technique [0002] With the development of the Internet, various social network platforms have appeared in the public's field of vision one by one. Whether it is Facebook abroad or Sina Weibo in China, the total number of users is increasing day by day. [1] . People not only watch and browse news hotspots and interesting stories on social networking platforms, but also participate in discussions and share their own lives, and a large amount of information circulates on social networking platforms. But it is precisely because of the existence of massive information and massive users that social networking platforms have attracted a large number of social robots [2,3] , they pretend to be normal users, publish advertisements and product links on social platforms, and some social robots...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more
Application Information
Patent Timeline
Application Date:The date an application was filed.
Publication Date:The date a patent or application was officially published.
First Publication Date:The earliest publication date of a patent with the same application number.
Issue Date:Publication date of the patent grant document.
PCT Entry Date:The Entry date of PCT National Phase.
Estimated Expiry Date:The statutory expiry date of a patent right according to the Patent Law, and it is the longest term of protection that the patent right can achieve without the termination of the patent right due to other reasons(Term extension factor has been taken into account ).
Invalid Date:Actual expiry date is based on effective date or publication date of legal transaction data of invalid patent.
Login to view more 
Login to view more