Keyword Obfuscation via Hash Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing keyword matching systems in computer monitoring are vulnerable to evasion by savvy users who can identify and avoid using keywords of interest, as these systems typically store keywords in accessible forms.
Innovation Solution
Generating a regular expression and hash value for each keyword, which are transmitted to a monitored computer to evaluate text without revealing the actual keywords, allowing for keyword matching without exposing the keyword list.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If keywords are stored in accessible forms for pattern matching, then keyword matching efficiency is improved, but users can identify and avoid using these keywords
Solution Approach 1:
The patent introduces hash values as an intermediary representation of keywords. Instead of storing and comparing actual keywords, the system stores hash values generated from keywords and compares hash values of monitored text against these stored hash values. This intermediary mechanism enables efficient matching while preventing users from identifying the actual keywords to avoid.
Solution Approach 2:
The patent transforms keywords from their original textual form into a different parameter representation (hash values). This parameter change maintains the essential matching capability while altering the form to be unintelligible to users, thus preventing evasion while preserving detection functionality.
2Reliability
If keywords are obfuscated to prevent user identification, then detection reliability is improved, but keyword matching complexity increases
Solution Approach 1:
The patent replaces the mechanical/textual comparison system with a mathematical hash-based system. Instead of comparing text strings directly, the system uses hash function algorithms to transform and compare representations, reducing the complexity of exact string matching while improving reliability through cryptographic-style obfuscation.
3Measurement precision
If actual keywords are transmitted to monitored computers, then matching accuracy is improved, but security is worsened as users can identify and avoid keywords
Solution Approach 1:
The patent extracts the essential matching property from keywords by generating hash values, then transmits only these extracted hash representations to monitored computers. The actual keywords remain secured at the source system, while the distributed systems receive only the unintelligible hash values needed for accurate matching without security vulnerabilities.
Data Source
AI summary
Embodiments of the invention provide novel systems, methods and software for searching and/or analyzing text, based on a list of one or more keywords of interest. In a set of embodiments, a first computer generates a regular expression satisfied by a particular keyword of interest as well as a hash value corresponding to the keyword. In some embodiments, the regular expression and the hash value are transmitted to a second computer, which identifies a set of text associated with a particular event and/or evaluates the set of text against the regular expression. If the text satisfies the regular expression, the second computer might generate a text hash from the text. This text hash can be compared with the hash from the package, and if there is a match, the event might be identified as requiring monitoring, collection and/or further analysis.