Nec corporation (20240095600). PRIVACY PRESERVING JOINT TRAINING OF MACHINE LEARNING MODELS simplified abstract

From WikiPatents
Jump to navigation Jump to search

PRIVACY PRESERVING JOINT TRAINING OF MACHINE LEARNING MODELS

Organization Name

nec corporation

Inventor(s)

Roberto Gonzalez Sanchez of Heidelberg (DE)

Vittorio Prodemo of Madrid (ES)

Marco Gramagiia of Madrid (ES)

PRIVACY PRESERVING JOINT TRAINING OF MACHINE LEARNING MODELS - A simplified explanation of the abstract

This abstract first appeared for US patent application 20240095600 titled 'PRIVACY PRESERVING JOINT TRAINING OF MACHINE LEARNING MODELS

Simplified Explanation

The abstract describes a method for training a shared machine learning model by generating a data transformation function, sharing it with other entities, creating private datasets, receiving private datasets from other entities, and training the model using the private datasets.

  • Explanation of the patent:

- Generating a data transformation function by a first entity - Sharing the data transformation function with one or more second entities - Creating a private dataset by applying the data transformation function to a first dataset of the first entity - Receiving private datasets from other entities, created by applying the data transformation function to their datasets - Training a machine learning model using the private datasets to produce a trained model

Potential applications of this technology: - Collaborative machine learning projects - Privacy-preserving machine learning applications - Federated learning systems

Problems solved by this technology: - Ensuring data privacy in shared machine learning projects - Allowing multiple entities to collaborate on training a model without sharing raw data - Improving model performance by training on diverse datasets

Benefits of this technology: - Enhanced data security and privacy - Increased collaboration opportunities in machine learning research - Improved model accuracy through training on diverse datasets

Potential commercial applications of this technology: - Secure data sharing platforms for machine learning projects - Collaborative research tools for data scientists and machine learning engineers - Privacy-focused machine learning services for businesses

Possible prior art: - Federated learning techniques in machine learning - Secure multi-party computation for collaborative data analysis

      1. Unanswered Questions:
        1. How does this method handle data synchronization issues between different entities in the training process?

The abstract does not provide details on how data synchronization is managed when training the shared machine learning model with private datasets from multiple entities.

        1. What measures are in place to prevent data leakage or unauthorized access to the shared machine learning model during the training process?

The abstract does not mention specific security measures or protocols implemented to safeguard the privacy and security of the shared machine learning model and private datasets.


Original Abstract Submitted

systems and method for training a shared machine learning (ml) model. a method includes generating, by a first entity, a data transformation function; sharing, by the first entity, the data transformation function with one or more second entities; creating a first private dataset, by the first entity, by applying the data transformation function to a first dataset of the first entity; receiving one or more second private datasets, by the first entity, from the one or more second entities, each second private dataset having been created by applying the data transformation function to a second dataset of the second entity; and training a machine learning (ml) model using the first private dataset and the one or more second private datasets to produce a trained ml model.