17540660. PRIVACY-PRESERVING CLASS LABEL STANDARDIZATION IN FEDERATED LEARNING SETTINGS simplified abstract (INTERNATIONAL BUSINESS MACHINES CORPORATION)
PRIVACY-PRESERVING CLASS LABEL STANDARDIZATION IN FEDERATED LEARNING SETTINGS
Organization Name
INTERNATIONAL BUSINESS MACHINES CORPORATION
Inventor(s)
Shonda Adena Witherspoon of Yorktown Heights NY (US)
Ramasuri Narayanam of Bangalore (IN)
Sameep Mehta of Bangalore (IN)
PRIVACY-PRESERVING CLASS LABEL STANDARDIZATION IN FEDERATED LEARNING SETTINGS - A simplified explanation of the abstract
This abstract first appeared for US patent application 17540660 titled 'PRIVACY-PRESERVING CLASS LABEL STANDARDIZATION IN FEDERATED LEARNING SETTINGS
Simplified Explanation
The patent application describes methods, systems, and computer program products for privacy-preserving class label standardization in federated learning settings.
- The method involves determining a signature for each class of data for multiple client devices in a federated learning environment using data privacy-preserving techniques.
- Signature matches are identified across the client devices, and class labels are generated for the matched data.
- The matched data is then labeled with the generated class labels across the client devices.
- Automated actions can be performed based on the labeled data.
Potential applications of this technology:
- Federated learning: The technology can be used in federated learning settings where multiple client devices collaborate to train a machine learning model without sharing their raw data.
- Privacy-preserving data analysis: The methods and systems described in the patent application enable privacy-preserving analysis of data across multiple client devices.
- Standardization of class labels: The technology provides a way to standardize class labels across different client devices, ensuring consistency in the analysis and interpretation of data.
Problems solved by this technology:
- Privacy concerns in federated learning: The methods and systems address privacy concerns by using data privacy-preserving techniques to determine signatures and match data without exposing sensitive information.
- Inconsistent class labels: The technology solves the problem of inconsistent class labels across different client devices by generating standardized labels based on signature matches.
Benefits of this technology:
- Privacy preservation: The methods and systems ensure that sensitive data remains private during the analysis process.
- Improved data analysis: By standardizing class labels, the technology improves the accuracy and reliability of data analysis across multiple client devices.
- Efficient collaboration: The technology enables efficient collaboration in federated learning settings by allowing client devices to share and analyze data without compromising privacy.
Original Abstract Submitted
Methods, systems, and computer program products for privacy-preserving class label standardization in federated learning settings are provided herein. A computer-implemented method includes determining, using one or more data privacy-preserving techniques, a signature for each of one or more classes of data for each of multiple client devices within a federated learning environment; identifying one or more signature matches across at least a portion of the multiple client devices; generating one or more class labels for the one or more classes of data associated with the one or more signature matches; labeling, across the at least a portion of the multiple client devices, the one or more classes of data associated with the one or more signature matches with the one or more generated class labels; and performing one or more automated actions based at least in part on the one or more labeled classes of data.
