Privacy-aware Event Detection via Primitive Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional security systems face challenges in detecting security events while preserving privacy, maintaining detection accuracy, reducing bias, and managing data efficiently, particularly due to limitations in handling sensitive information and sparse training data.
Innovation Solution
A privacy-aware event detection method that extracts primitive values from raw measurements, excluding personally identifiable information and sensitive attributes, and uses these anonymized values for event detection, enabling centralized and complex event models, reduced data storage, and faster analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If raw measurements including personally identifiable information are used for event detection, then detection accuracy may be improved, but privacy is compromised
Solution Approach 1:
The system extracts and removes personally identifiable information and sensitive attributes from raw measurements before event detection. Only anonymized primitive values are retained and transmitted to the event detection system, effectively extracting the harmful PII components while preserving useful event detection capabilities.
Solution Approach 2:
A primitive extraction system acts as an intermediary between raw measurements and event detection. This intermediary layer transforms raw data containing PII into anonymized primitive values, serving as a mediator that protects privacy while enabling accurate event detection through structured, privacy-preserving data representation.
2Object-affected harmful factors
If anonymized primitive values are used instead of raw measurements, then privacy is preserved and data storage is reduced, but detection accuracy may deteriorate
Solution Approach 1:
The system segments raw measurements into distinct primitive components (e.g., object presence, motion detection, temporal patterns) that are necessary for event detection. This segmentation organizes data into structured, meaningful units that preserve detection accuracy while enabling privacy protection through selective anonymization of each primitive.
Solution Approach 2:
The system transforms raw measurements by changing their parameter representation from detailed, identifiable data to structured primitive values with controlled granularity. This parameter transformation maintains the essential features needed for event detection while altering the data form to eliminate personally identifiable information.
3Adaptability or versatility
If centralized event detection systems are used, then model complexity and training data aggregation are improved, but data transmission and processing latency increase
Solution Approach 1:
The system performs preliminary processing by extracting primitive values and performing initial event detection locally before transmitting results to the centralized system. This preliminary action reduces the amount of data that needs to be transmitted and processed centrally, thereby reducing latency while still enabling access to complex centralized models when needed.
Data Source
AI summary
Variants of the privacy-aware event detection method can include: determining measurements of a monitored space, determining values for a set of primitives from the measurements, optionally sending the primitive values to a remote event detection system, detecting an event based on the primitive values, optionally analyzing the event, and optionally notifying a user of a security threat.