17743906. METHOD AND APPARATUS WITH NEURAL NETWORK ARCHITECTURE SEARCH simplified abstract (SAMSUNG ELECTRONICS CO., LTD.)

From WikiPatents
Jump to navigation Jump to search

METHOD AND APPARATUS WITH NEURAL NETWORK ARCHITECTURE SEARCH

Organization Name

SAMSUNG ELECTRONICS CO., LTD.

Inventor(s)

WONHEE Lee of Yongin-si (KR)

METHOD AND APPARATUS WITH NEURAL NETWORK ARCHITECTURE SEARCH - A simplified explanation of the abstract

This abstract first appeared for US patent application 17743906 titled 'METHOD AND APPARATUS WITH NEURAL NETWORK ARCHITECTURE SEARCH

Simplified Explanation

The disclosed patent application describes a method and apparatus for finding the best architecture for a neural network. Here are the key points:

  • The apparatus includes a processor that generates a neural network loss based on the parameters of a candidate architecture for the neural network.
  • It measures the first hardware resources used in operating the neural network with the candidate architecture.
  • Using a hardware resource prediction model, it generates a prediction of the second hardware resources that would be used for operating the neural network with the candidate architecture.
  • It determines a hardware resource loss based on the comparison between the first and second hardware resources.
  • Finally, it determines the optimal architecture of the neural network based on the neural network loss and the hardware resource loss.

Potential applications of this technology:

  • This technology can be used in various fields where neural networks are employed, such as image recognition, natural language processing, and autonomous vehicles.
  • It can help researchers and developers in designing and optimizing neural networks for specific tasks, improving their performance and efficiency.

Problems solved by this technology:

  • Finding the optimal architecture for a neural network can be a time-consuming and resource-intensive process.
  • This technology automates the process of searching for the best architecture, saving time and effort for researchers and developers.

Benefits of this technology:

  • By automatically searching for the optimal architecture, this technology can significantly reduce the time and resources required for designing and optimizing neural networks.
  • It can lead to improved performance and efficiency of neural networks, resulting in better accuracy and faster processing times.


Original Abstract Submitted

Disclosed is a method and apparatus for searching for an optimal architecture of a neural network. The apparatus may include a processor configured to generate a neural network loss based on parameters of a candidate architecture for the neural network, measure first hardware resources used in operation of the neural network with the candidate architecture, generate a prediction, using a hardware resource prediction model, of second hardware resources that would be used for operating the neural network with the candidate architecture, determine a hardware resource loss based on the first hardware resources and the second hardware resources, and determine a target architecture of the neural network based on the neural network loss and the hardware resource loss.

Retrieved from "http://wikipatents.org/index.php?title=17743906._METHOD_AND_APPARATUS_WITH_NEURAL_NETWORK_ARCHITECTURE_SEARCH_simplified_abstract_(SAMSUNG_ELECTRONICS_CO.,_LTD.)&oldid=25525"