Search Query Obfuscation Using AI-Generated Decoy Queries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for obfuscating search queries are inadequate, as they either fail to conceal user identity or result in easily identifiable decoy queries, leading to potential exposure of sensitive information.
Innovation Solution
Utilizing a combination of hierarchical term classes and generative language models to generate decoy queries that are harder to identify by organizing input queries into known terms and determining classifications, then generating decoy queries based on complexity and similarity scores.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If fixed sets of other terms are used to generate decoy queries, then the obfuscation process is simple, but the decoy queries are easily identifiable
Solution Approach 1:
The patent changes the parameters of decoy query generation by using generative language models instead of fixed term sets. The model generates decoys based on multiple factors including query classification, semantic similarity, and complexity scoring, transforming the static generation process into a dynamic one that produces more indistinguishable decoys
Solution Approach 2:
The patent replaces the mechanical/simple approach of using fixed term sets with a computational/algorithmic approach using generative language models. This substitution enables the system to generate context-aware, semantically similar decoys that are much harder to identify as fake queries
2Device complexity
If simple obfuscation methods are used, then the system complexity is low, but user identity and sensitive information are exposed
Solution Approach 1:
The patent introduces an intermediary layer in the form of a generative language model that processes the original query and transforms it into obfuscated decoys. This intermediary component enables strong obfuscation while maintaining system manageability through modular architecture
Solution Approach 2:
The system performs preliminary classification of the input query into hierarchical categories before generating decoys. This preliminary action organizes the query structure and enables more effective obfuscation by tailoring the decoy generation to the specific query context and classification
3Productivity
If decoy queries are generated without classification, then the generation process is fast, but the obfuscation effectiveness is reduced
Solution Approach 1:
The patent performs query classification into hierarchical categories as a preliminary step before decoy generation. This classification organizes the query structure and enables more effective obfuscation by tailoring the decoy generation to the specific query context, thereby improving reliability without significantly impacting speed
Data Source
AI summary
Systems, methods, and computer program products for obfuscating search queries are described herein. A method comprises reading an input query; reading a set of known terms organized into hierarchical classes; determining whether the input query is included in the set of known terms; determining one or more classifications for the input query in accordance with its inclusion in the set of known terms; generating a prompt in accordance with the one or more classifications and the input query; providing the prompt to a generative language model as input; receiving, from the generative language model, a plurality of candidate queries in accordance with the prompt; determining a score for the plurality of candidate queries; and generating a plurality of decoy queries based on the plurality of candidate queries and the score.