Edge User Profiling With On-Device Anonymization for Private Communication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing communication systems fail to adequately protect user privacy, secure user data, and efficiently utilize AI/ML workloads, leading to data breaches, loss of privacy, and high computing costs.
Innovation Solution
Implementing a method and system where edge devices store and process user data, assign tags and scores, and share anonymized user IDs with cloud servers, maintaining personal data security and enabling cost-effective AI/ML workloads on the edge.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If user data is stored and processed on cloud servers, then centralized data processing capability is improved, but user data security and privacy protection deteriorate
Solution Approach 1:
The patent segments the data processing architecture into edge devices (local processing) and cloud servers (centralized coordination). User data remains stored and processed locally on edge devices, while only anonymized identifiers and aggregated insights are transmitted to the cloud. This segmentation maintains data security at the source while enabling centralized coordination benefits.
Solution Approach 2:
The patent introduces anonymized user identifiers as an intermediary between user data and cloud processing. Instead of transmitting actual user data to the cloud, the system uses anonymized IDs that allow the cloud server to coordinate processing and deliver contextual content without having access to or storing sensitive personal information, thus maintaining a security buffer.
2Measurement precision
If all user data is transmitted to cloud servers for processing, then comprehensive data analysis capability is improved, but data breach risks and computing costs worsen
Solution Approach 1:
The patent extracts only the essential elements needed for cloud processing (anonymized user IDs and aggregated activity patterns) while leaving the bulk of sensitive user data on edge devices. This extraction approach enables the cloud server to perform coordination and contextual content delivery without requiring access to comprehensive user data, thereby reducing data breach risks while maintaining analysis capability.
Solution Approach 2:
The patent performs preliminary data processing and anonymization at the edge device before any data leaves the local environment. User activities are processed, anonymized, and aggregated locally first, then only the processed results are transmitted to the cloud. This preliminary action ensures data security is maintained from the outset while enabling effective cloud-based coordination.
3Power
If AI/ML workloads are processed on cloud servers, then computing power is improved, but computing costs and energy consumption worsen
Solution Approach 1:
The patent segments AI/ML workloads between edge devices and cloud servers based on computational requirements. Heavy data processing and model inference occur locally on edge devices using collected user activity data, while the cloud server handles only lightweight coordination tasks and contextual content delivery. This segmentation reduces cloud computing costs and energy consumption while maintaining necessary processing capabilities.
Solution Approach 2:
The patent enables edge devices to perform self-service AI/ML processing by running local models that process user activity data and generate insights without requiring constant cloud intervention. The edge device autonomously processes data, generates anonymized identifiers, and only engages the cloud server when contextual content delivery is needed, thereby reducing overall system computing costs and energy consumption.
Data Source
AI summary
Present disclosure provides a method and a system for privacy-preserving communication on wireless edge devices. The proposed method includes creating and maintaining user data (user activities) on the wireless edge devices. Further, the method includes running on-device mechanism on the user data to derive the daily routine of a user, and assigning a score to multiple activities (walking, running, climbing the stairs, descending the stairs etc.) in the daily routine. Further, the method includes creating and maintaining a profile of the user based on these activities. Further, the user profile is used by the system as a basis for targeted and contextual communication with the user in a manner that preserves user personally identifiable and sensitive information on the wireless edge device.