Peer-Based Anomalous Rights Detection Using Vector Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large institutions face challenges in tracking and managing access privileges as employees change roles, making it difficult to determine anomalous rights, which can be exploited by fraudulent actors for unauthorized access to sensitive information.
Innovation Solution
A peer-based anomalous rights detection system uses rights vectors and nearest neighbor algorithms to identify individuals with anomalous access rights by comparing their rights profiles to those of peers, employing techniques like Approximate Nearest Neighbors Oh Yeah (ANNOY) for efficient computation and calculating P scores and Z-scores to flag potentially unauthorized access.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If manual tracking of access privileges is used in large institutions, then administrative overhead is reduced, but the ability to detect anomalous rights deteriorates
Solution Approach 1:
The system automatically computes P-scores and Z-scores for each individual by comparing their rights vector to peer groups, eliminating the need for manual tracking while enabling precise detection of anomalous rights through self-performed statistical analysis
Solution Approach 2:
The patent replaces manual administrative tracking with an automated computational system that uses nearest neighbor algorithms and statistical scoring (P-scores and Z-scores) to detect anomalous rights, substituting human effort with algorithmic processing
2Measurement precision
If detailed tracking of all individuals' rights is implemented, then detection of anomalous rights improves, but system complexity worsens
Solution Approach 1:
The system segments the population into peer groups based on role categories and computes statistics independently for each segment, allowing detailed detection within groups while maintaining overall system manageability through modular organization
Solution Approach 2:
The patent creates a universal framework that handles diverse rights types and individual roles through a common statistical approach (P-scores and Z-scores), enabling the system to manage complexity by applying the same methodology across different domains
3Reliability
If real-time monitoring of rights changes is performed, then detection capability improves, but computational load worsens
Solution Approach 1:
The system pre-computes peer group statistics and P-scores for all individuals, establishing a baseline before actual monitoring begins, which reduces the computational burden during real-time operation by eliminating the need for repeated full-scale comparisons
Solution Approach 2:
The patent implements continuous monitoring that computes Z-scores based on deviations from established peer group norms, providing feedback when anomalies are detected and enabling reliable real-time detection without requiring exhaustive re-computation of all statistics
Data Source
AI summary
Devices and techniques are generally described for peer-based anomalous rights detection. In various examples, a rights vector may be determined for a first individual, the rights vector representing rights held by the first individual. A nearest neighbor algorithm may be used to determine a set of individuals having similar rights to the first individual. In various examples, a category label associated with the first individual may be determined. In some examples, a number of individuals of the set of individuals having the category label may be determined. In some examples, a determination may be made that the rights held by the first individual are anomalous based at least in part on the number. In some cases, alert data indicating that the rights held by the first individual are anomalous may be generated.