Open Access
Review
Table 2.
Summary of behavioral conformity authentication
| Object | Description | Characteristics | Issues |
|---|---|---|---|
| Fraud [5, 23, 101–107] | Fraud refers to the behavior of fraudsters for obtaining illegal benefits through fraudulent means, such as concealing facts and stealing information. | (1) Ensuring process security by detecting suspicious activity during transactions. (2) Uncovering hidden fraud networks through behavioral correlation analysis. | (1) Possible leakage and illegal use of user privacy data. (2) Imbalanced training data. (3) Hard to deal with new fraud patterns timely. |
| Malicious intrusion [6, 8, 108–116] | Malicious intrusion refers to attacks from outside the system, posing threats to personal computers, mobile phones and even the entire network. | (1) Ensuring system security by detecting malicious activity. (2) Timely alerts and responses. | (1) Difficulty in detecting advanced persistent threats. (2) Imbalanced training data. |
| Insider threat [7, 117–122] | Insider threat refers to the risks existing in the system, affecting the safety of the system from the inside. | (1) Enhancing network safety and system availability. (2) Strengthening the compliance of systems, such as enterprise networks. | (1) Dependencies on high-quality behavioral data. (2) Difficult to address zero-day backdoors. |
| Unfair discrimination [123–128] | Unfair discrimination refers to unfairness or imbalance within a system, which gradually become apparent as the system runs. | Detecting patterns of unfair discrimination at an early stage allows system personnel to intervene and correct such practices timely, which avoids the cumulative risk of unfair discrimination. | The scarcity of data complicates the comprehensive selection of fairness indicators and hinders the establishment of universally applicable fairness metrics. |
| Privacy leakage [129–136] | Privacy leakage usually refers to the exposure of users’ private data due to security issues, but it also includes disclosure by a second party, a third party, or even other users. | (1) The discovery of external threats and enhancement of system security through privacy leak detection. (2) The improvement of privacy safety through privacy protection measures. | Under data constraints, such as scenarios where there are no clearly defined levels of privacy data, establishing an efficient detection model is a subsequent challenge. |
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.