With the in-depth development of AI technology, "AI face swapping" and "AI stripping" tools have become common. Some lawbreakers have turned their crooked ideas into creating "pornographic rumors" and even threatening and defrauding them. On April 19, the Haidian District People's Court in Beijing heard a case of using "AI stripping" software to produce obscene materials. The defendant used deepfake technology and AI software to create nude pictures of women and sell them for profit, and at the same time sold "AI clothes removal" software and tutorials for profit. (quoted from "Southern Metropolis Daily" on May 13)
In today's society, there are frequent online rumor incidents, and while we are vigilant against AI rumors, we should also pay more attention to big data anti-pornography. In this article, we will briefly analyze the principle of SNA algorithms that are common in police intelligence analysis systems.
1. What is the SNA algorithm?
In simple terms, the SNA algorithm is called a social network algorithm. Social network analysis (SNA) is a branch of complex systems that can be used to build multi-intelligence systems.
Firstly, while studying how social network analysis can assist in the modeling of multi-intelligent systems, it is necessary to address the similarities and differences between the two examples. Therefore, before constructing the SNA system, it is necessary to construct a prototype of a multi-intelligent system system, and solve the task by forming an intelligent cluster based on the social network established between the agent instances.
Let's take a look at the definition of MAS, which is defined as a model that represents social structures (such as organizations and alliances) in order to objectively analyze the associative behavior of open systems. Organizations and alliances are made up of individuals who are interconnected through different types of relationships (e.g., dependence on goals, conflicts over resources, similar beliefs, etc.).
After clarifying the definitions of SNA and MAS, we can see that there is a close relationship between multi-agent systems (MAS) and complex systems theory, and they can be represented in the context of social network analysis. The combination of MAS and SNA technology in the principle of intelligence analysis system enables the agent society to use SNA indicators in the decision-making process and consider the attributes such as the location, value and importance of agents in the network.
2. The role of SNA
Social network analysis (SNA) is a field of scientific research that originates from fields such as sociology, social psychology, and anthropology. This field studies the links between social actors (also known as relational bonds).
In SNA, roles can be individuals or companies, and can be analyzed as individual units or as collective social units.
The fundamental difference between SNA and other studies is that it does not focus on the attributes (characteristics) of actors, but rather on the connections between them. The relationship between actors is established by relational bonds or connections.
The most common types of connections are: individual assessments (e.g., friendships or emotional exchanges); transactions and transfers of material resources (purchase and sale transactions between two companies); transfer of immaterial resources (exchange of electronic messages); When the participants of the behavior and the event occur at the same time, and have a regional, behavior, etc., correlation or affiliation, then the transaction network is formed. For example, network association cluster relationship (social network connection between QQ, Weibo, and WeChat users); the connection between formal nodes (the cycle of authority between superiors and subordinates in the company); biological relationships (paternity) and so on.
SNA is a method of enhancing knowledge sharing by analyzing the location and structure between actors, i.e., their relationships.
Then you might be asking
- This paper mainly discusses the SNA-based inference method composed of multi-agent teams, so how can SNA be applied to the decision-making process of multi-agent systems?
- In this paper, we mentioned the close connection between multi-agent systems and social networks, so how to use SNA indicators to evaluate the status and importance of agents in the network in this integration process?
So let's talk about the practical utility of social networks through an example of this.
3. Cases of application fields of SNA
In the field of intelligence analysis systems, the IBM I2 we are familiar with is a commonly used police intelligence analysis system. The underlying logic of IBM I2 uses the model application of SNA.
Let's take a look at an example:
Picture excerpt from-360doc [Knowledge] IBM i2 Visual Analytics Product Overview (Network Diagram Invasion and Deletion)
In this case, through the network node analysis of IBM i2, the relationship cluster calculation of the associated node is derived from the central node. The prototype uses SNA principles to demonstrate a model of dynamic formation of teams, where nodes are formed spontaneously and dispersively after decisions based on their local social networks. The participants of the cluster have a social network.
For a node to be part of a team, it must have a social connection (i.e., an edge in the network) with at least one member of the team. Thus, in this cluster, the multi-node contains N sub-nodes, A={a1,a2,...,aN}, where each sub-node can be considered as a node in the social network. The network is modeled as an adjacency matrix E, where each adjacency matrix element, eij=1, if present, connects the edges of the two surrogate ai and aj, otherwise eij=0. To put it simply, it is a 0 out boundary. That is, the last node in the boundary. Then the selected scope of the cluster is determined.
IV. Conclusion
After graduating from university, the author worked as a new media reporter in a national legal journal for two years, and then devoted himself to the IT field to engage in the design of information analysis system products in the field of risk control for many years.
The author would like to say: in the current era of AI pandemic, we must be aware of such a point of view from the perspective of risk control.
The more advanced science and technology are, the higher the possibility of spreading rumors on the Internet and people with ulterior motives violating the law online.
But it's still an old saying: Skynet is magnificent but not leaky.
The progress of the times is also the progress of the rule of law on the Internet. There will be more high-tech systems to eliminate, detect, and arrest those who use AI and the Internet to spread rumors, fraud, and illegal acts. Don't leave anything to chance. Respect science and technology, abide by good morals, and be a good citizen.
This article was originally published by @kingwu on Everyone is a Product Manager. Reproduction without the permission of the author is prohibited
The title image is from Unsplash and is licensed under CC0
The views in this article only represent the author's own, everyone is a product manager, and the platform only provides information storage space services