18577549. TRANSFER LEARNING METHOD FOR A MACHINE LEARNING SYSTEM simplified abstract (Eaton Intelligent Power Limited)
Contents
TRANSFER LEARNING METHOD FOR A MACHINE LEARNING SYSTEM
Organization Name
Eaton Intelligent Power Limited
Inventor(s)
Tony Marrero of Dublin 7, Dublin (IE)
Catriona Clarke of Kildare, Newbridge (IE)
Adi Botea of Clonsilla, Dublin (IE)
Chahrazed Bouhini of Dublin 1, Dublin (IE)
Suchandra Mandal of Galway (IE)
TRANSFER LEARNING METHOD FOR A MACHINE LEARNING SYSTEM - A simplified explanation of the abstract
This abstract first appeared for US patent application 18577549 titled 'TRANSFER LEARNING METHOD FOR A MACHINE LEARNING SYSTEM
Simplified Explanation:
This patent application describes a transfer learning method for a system with existing agents and a new agent, where the new agent needs to model a new scenario. The method involves selecting an available model from a database and using transfer learning data to train the selected model based on the new scenario metadata and training data.
Key Features and Innovation:
- Transfer learning method for a system with existing agents and a new agent
- Selection of an available model from a database to model a new scenario
- Training the selected model using transfer learning data based on the new scenario metadata and training data
Potential Applications: This technology can be applied in various fields such as healthcare, finance, and autonomous systems where new scenarios need to be modeled efficiently.
Problems Solved: This technology addresses the challenge of efficiently modeling new scenarios in a system with existing agents and trained machine learning models.
Benefits:
- Faster and more accurate modeling of new scenarios
- Utilization of existing trained models for transfer learning
- Improved performance of the new agent in modeling new scenarios
Commercial Applications: Potential commercial applications include predictive analytics, anomaly detection, and personalized recommendations in various industries.
Prior Art: Prior art related to this technology may include research on transfer learning methods in machine learning systems and databases of machine learning models for different scenarios.
Frequently Updated Research: Researchers are constantly exploring new techniques and algorithms for transfer learning in machine learning systems, which could further enhance the efficiency and effectiveness of this technology.
Questions about Transfer Learning: 1. How does transfer learning differ from traditional machine learning methods? Transfer learning leverages knowledge from existing models to improve learning in new scenarios, while traditional machine learning starts from scratch for each new task.
2. What are some common challenges in implementing transfer learning in real-world applications? Some challenges include domain adaptation, dataset bias, and selecting the most appropriate pre-trained model for transfer learning.
Original Abstract Submitted
A transfer learning method for a system including a plurality of existing agents each including a trained machine learning model for modelling a respective existing machine learning scenario, and a new agent. The system includes a database comprising available models for modelling scenarios, including the trained models, existing scenario metadata indicative of existing scenarios, and transfer learning data indicative of parts of the trained models. The method comprises receiving new scenario metadata indicative of a new scenario to be modelled by the new agent, and receiving new scenario training data for training a model of the new scenario. The method also includes querying the database to: select an available model, based on the received data, to model the new scenario; and, select at least some of the transfer learning data, based on the received data, to train the selected model.
