View source for 18659457. ELECTRONIC APPARATUS AND METHOD FOR RE-LEARNING TRAINED MODEL simplified abstract (SAMSUNG ELECTRONICS CO., LTD.)

=ELECTRONIC APPARATUS AND METHOD FOR RE-LEARNING TRAINED MODEL=


==Organization Name==

[[:Category:SAMSUNG ELECTRONICS CO., LTD.|SAMSUNG ELECTRONICS CO., LTD.]]


[[Category:SAMSUNG ELECTRONICS CO., LTD.]]

==Inventor(s)==

[[:Category:Sungju Hwang of Daejeon (KR)|Sungju Hwang of Daejeon (KR)]][[Category:Sungju Hwang of Daejeon (KR)]]

[[:Category:Jaehong Yoon of Daejeon (KR)|Jaehong Yoon of Daejeon (KR)]][[Category:Jaehong Yoon of Daejeon (KR)]]

[[:Category:Jeongtae Lee of Busan (KR)|Jeongtae Lee of Busan (KR)]][[Category:Jeongtae Lee of Busan (KR)]]

[[:Category:Eunho Yang of Daejeon (KR)|Eunho Yang of Daejeon (KR)]][[Category:Eunho Yang of Daejeon (KR)]]

==ELECTRONIC APPARATUS AND METHOD FOR RE-LEARNING TRAINED MODEL - A simplified explanation of the abstract==


This abstract first appeared for US patent application 18659457 titled 'ELECTRONIC APPARATUS AND METHOD FOR RE-LEARNING TRAINED MODEL

**Simplified Explanation:**
The patent application describes a method for re-learning a trained model by selectively re-learning parameters associated with a new task while dynamically expanding the size of the trained model if needed.

**Key Features and Innovation:**
* Re-learning a trained model by identifying neurons associated with a new task and selectively re-learning parameters for those neurons.
* Dynamically expanding the size of the trained model to reconstruct the input model if a preset loss value is reached.
* Enhancing the efficiency of re-learning tasks in neural networks by focusing on specific neurons and parameters.

**Potential Applications:**
This technology could be applied in various fields such as machine learning, artificial intelligence, data analysis, and pattern recognition.

**Problems Solved:**
* Improves the adaptability of trained models to new tasks.
* Enhances the efficiency of re-learning processes in neural networks.
* Allows for targeted parameter adjustments in complex models.

**Benefits:**
* Faster adaptation to new tasks.
* Improved accuracy in re-learning processes.
* Enhanced performance of neural networks in dynamic environments.

**Commercial Applications:**
Potential commercial applications include automated data analysis systems, intelligent decision-making algorithms, and advanced pattern recognition software for various industries.

**Prior Art:**
Prior research in the field of neural network re-training and dynamic model expansion could provide valuable insights into similar approaches.

**Frequently Updated Research:**
Stay updated on the latest advancements in neural network re-training techniques, dynamic model expansion methods, and optimization strategies for complex models.

**Questions about the Technology:**
1. How does the method for selectively re-learning parameters in neural networks improve efficiency?
2. What are the potential implications of dynamically expanding the size of trained models in machine learning applications?


==Original Abstract Submitted==

A method for re-learning a trained model is provided. The method for re-learning a trained model includes: receiving a data set including the trained model consisting of a plurality of neurons and a new task; identifying a neuron associated with the new task among the plurality of neurons to selectively re-learn a parameter associated with the new task for the identified neuron; and dynamically expanding a size of the trained model on which the selective re-learning is performed if the trained model on which the selective re-learning has a preset loss value to reconstruct the input trained model.



[[Category:G06N3/08]]
[[Category:G06N3/04]]
[[Category:G06N3/045]]
[[Category:G06N3/082]]
[[Category:CPC_G06N3/08]]