Cloud File PII Detection Through Sampling and Data Standardization

The patent applies partial action by processing only a sample subset of files rather than the entire dataset. The system selects a representative sample of files from the cloud storage system and processes only those samples through the PII detection pipeline, achieving acceptable detection accuracy while dramatically reducing processing time and computational resources compared to exhaustive processing of all files.

The system processes only a sampled subset of files rather than the complete dataset, reducing computational resource consumption and energy usage while maintaining acceptable detection accuracy. This partial processing approach avoids the excessive resource demands of analyzing every file in the cloud storage system.

The patent creates and processes copies of PII data rather than exposing the actual original data. The system generates synthetic PII data that mimics the structure and characteristics of real PII for training and testing purposes, and processes sampled copies during detection operations. This copying approach enables accurate detection model development and operation while maintaining data security and compliance by never exposing the actual sensitive PII stored in the cloud system.

The patent segments the diverse file types into distinct categories (structured files, semi-structured files, unstructured files) and applies specialized processing methods tailored to each category. The system identifies the file type and routes it to the appropriate processing pipeline, enabling efficient handling of each file format while maintaining versatility across multiple file types. This segmentation avoids the inefficiency of applying a single uniform processing approach to all file types.