Session Data Semantic Analysis via Vector Embeddings
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing application performance monitoring (APM) technologies face challenges in comparing and analyzing session data from distributed computing systems due to its inhomogeneity and large size, making it difficult to identify fraudulent sessions and optimize user experiences.
Innovation Solution
A computer-implemented method that involves receiving session data, generating a textual description, creating a vector embedding using a text embedding model, and storing this embedding in a database, allowing for semantic analysis and comparison of session data. This method also includes identifying fraudulent sessions by comparing new session data with reference fraudulent data and synthetically testing websites using similar vector embeddings.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If session data is collected and stored in its original form, then complete session information is preserved, but data processing and comparison become difficult due to inhomogeneity and large size
Solution Approach 1:
The patent extracts the essential semantic information from raw session data by generating textual descriptions that capture the core meaning of user sessions. This extraction process separates the valuable semantic content from the voluminous raw data, making it manageable for comparison and analysis while preserving the essential information needed for fraud detection and user experience optimization.
Solution Approach 2:
The patent transforms session data from its original heterogeneous format into a standardized vector embedding representation. This parameter transformation converts complex, inhomogeneous session data into a uniform numerical format that enables efficient mathematical operations, similarity calculations, and database storage while maintaining the semantic meaning of the original data.
2Productivity
If vector embeddings are generated for all session data, then semantic comparison becomes efficient, but storage requirements increase
Solution Approach 1:
The patent creates compact vector embedding copies of the essential session characteristics rather than storing all original raw data. These vector representations serve as efficient proxies that capture the semantic meaning of sessions in a condensed format, enabling rapid comparison and analysis while significantly reducing the storage burden compared to keeping complete raw session data.
3Reliability
If traditional session data analysis methods are used, then all raw data can be examined, but fraudulent sessions are difficult to identify due to data inhomogeneity
Solution Approach 1:
The patent transforms heterogeneous session data into a unified vector embedding space where fraudulent sessions can be reliably detected through similarity comparisons. This parameter transformation enables the use of efficient distance-based algorithms to identify anomalous patterns that would be difficult to detect in the original inhomogeneous data format, significantly improving fraud detection accuracy.
Data Source
AI summary
The disclosure concerns a computer-implemented method for semantically analyzing session data, a computer-implemented method for identifying fraudulent session data captured in a distributed computing system, and a computer-implemented method for synthetically testing a website. The objective of the disclosure is to propose a similarity measure for session data in order to compare different sessions, such as user sessions or business process journeys, to each other. Another objective of the disclosure is to semantically analyze session data by similarity and to store the analysis result in a database. The objective is solved by receiving session data from a session occurring in the distributed computing system; generating a textual description for the session data; generating a vector embedding from the textual description, where the vector embedding represents the session data; and storing the vector embedding, along with a reference to the session data, in a database.