Bifurcated Mobile Data Sanitization for Leakage Prevention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data sanitization systems face challenges in identifying sensitive data due to the large size of regex routines, limited storage on mobile devices, and the lack of insight into how locally run artificial intelligence models sanitize data, leading to security risks and inefficiencies.
Innovation Solution
A bifurcated sanitization process using a machine learning model on the mobile device for initial sanitization, followed by remote server scrubbing, along with on-demand model updates and unconventional output formats to ensure accuracy and control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If regex routines are used to identify sensitive data, then data sanitization can be performed, but the large size of the regex corpus exceeds mobile device storage capacity
Solution Approach 1:
The patent segments the sanitization system into two parts: a lightweight local component on the mobile device that performs initial sanitization without requiring large regex corpora, and a remote server component that handles complex pattern matching. This segmentation allows the mobile device to maintain sanitization functionality while avoiding the storage burden of a complete regex corpus.
Solution Approach 2:
The patent introduces an intermediary machine learning model that acts as a mediator between the mobile device and the remote server. The model is trained to identify sensitive data patterns locally, serving as an intermediary that reduces the need for direct communication with the server and eliminates the requirement for storing large regex corpora on the device.
2Reliability
If data is transmitted to a remote server for regex processing, then sanitization instructions can be obtained, but security risks increase due to potential data exposure during transmission
Solution Approach 1:
The patent applies preliminary action by training the machine learning model in advance on the remote server. The trained model is then downloaded to the mobile device, where it performs sanitization locally without requiring further data transmission. This preliminary training action eliminates the need to transmit sensitive data to the server during operation, thereby reducing security risks.
Solution Approach 2:
The machine learning model enables the mobile device to perform sanitization autonomously using locally stored trained parameters. The device serves itself by making local classification decisions without requiring continuous communication with the remote server, thereby minimizing data exposure opportunities while maintaining sanitization accuracy.
3Adaptability or versatility
If multiple regex routines are implemented to detect different types of sensitive data, then comprehensive sanitization is achieved, but device complexity and storage requirements increase
Solution Approach 1:
The patent applies universality by designing a single machine learning model that can detect multiple types of sensitive data (PII, financial information, health records, etc.) through various input modalities (text, images, audio). This universal model replaces the need for multiple specialized regex routines, thereby maintaining comprehensive detection coverage while reducing system complexity.
Solution Approach 2:
The patent utilizes parameter changes by training the machine learning model with diverse training data representing different sensitive data types and formats. The model learns to adapt its detection parameters based on the input data characteristics, enabling it to handle various sensitive data types with a single unified system rather than requiring separate routines for each data type.
4Object-affected harmful factors
If a machine learning model is run locally on the mobile device, then data security is improved by preventing data transmission, but the lack of remote oversight reduces governance and control
Solution Approach 1:
The patent implements feedback by establishing a communication channel between the mobile device and remote server that allows the server to receive information about sanitization operations and provide updates. The server can monitor model performance, receive feedback on detection accuracy, and push model updates to devices, thereby maintaining governance and control while the model operates locally for security.
Solution Approach 2:
The patent applies preliminary action by having the remote server prepare and validate machine learning models before deploying them to mobile devices. The server performs model training, validation, and approval in advance, ensuring governance and control over the sanitization process while allowing local execution for security. This preliminary governance action maintains oversight without requiring continuous data transmission.
Data Source
AI summary
Methods and systems are described herein for leveraging artificial intelligence to sanitize sensitive data and prevent the data from leaving the mobile device and/or be exposed to unauthorized third parties. More specifically, methods and systems are described for a novel and unconventional architecture for a data sanitization application, a novel and unconventional delivery format for the data sanitization model, and a novel and unconventional output format of the data sanitization model.