18623569. FEDERATED LEARNING simplified abstract (Nokia Technologies Oy)

FEDERATED LEARNING

Organization Name

Nokia Technologies Oy

Inventor(s)

Soumyajit Chatterjee of Cambridge (GB)

Akhil Mathur of Cambridge (GB)

FEDERATED LEARNING - A simplified explanation of the abstract

This abstract first appeared for US patent application 18623569 titled 'FEDERATED LEARNING

Simplified Explanation: The patent application relates to federated learning for computational models, where an apparatus determines if a model can be trained locally on a client device within a target time. If not, a modified version of the model is selected for training within the target time.

Key Features and Innovation:

• Apparatus determines if a computational model can be trained locally within a target time.
• Selects a modified version of the model if local training is not feasible.
• Provides the modified model for local training on the client device.

Potential Applications: This technology can be applied in various fields such as healthcare, finance, and manufacturing where privacy concerns or limited network connectivity are present.

Problems Solved: The technology addresses the challenge of training complex computational models locally on client devices within a specified time frame.

Benefits:

• Enhances privacy by allowing local training of models on client devices.
• Improves efficiency by selecting modified models for faster training.
• Enables training in low-connectivity environments.

Commercial Applications: Potential commercial applications include secure healthcare data analysis, real-time financial forecasting, and on-device personalized recommendations in e-commerce.

Prior Art: Readers can explore prior research on federated learning, computational model training, and privacy-preserving machine learning techniques.

Frequently Updated Research: Stay informed about advancements in federated learning algorithms, model optimization techniques, and privacy-enhancing technologies in machine learning.

Questions about Federated Learning for Computational Models: 1. How does federated learning differ from traditional centralized machine learning approaches? 2. What are the key challenges in implementing federated learning for computational models?

Original Abstract Submitted

Example embodiments relate to an apparatus, method and computer program relating to federated learning for computational models. In an example, an apparatus comprises means for determining, based on one or more resources of a client device, whether a first computational model architecture can be trained locally by the client device within a target training time. The apparatus may also comprise means for selecting, if the first computational model architecture cannot be trained locally by the client device within the target training time, a modified version of the first computational model architecture that can be trained by the client device within the target training time. The apparatus may also comprise means for providing the selected modified version of the first computational model architecture for local training by the client device.