Wolfram U

Neural Networks Basic Concepts

Estimated Time: 17 min

Course Level: Beginner

Summary

Build and train your own convolutional neural network with the help of this video class. Wolfram Language provides an integrated high-level way to interface with neural nets that enables you to avoid low-level programming and focus on building sophisticated networks symbolically, and the video demonstrates how you can leverage this for your application. Examples show how to efficiently train networks on large out-of-core datasets and easily import and export them for use on all platforms. The class starts at a beginner level with a quick review of basic concepts like tensors and layers and goes on to illustrate how to train an entire network.

Featured Products & Technologies: Wolfram Language (available in Mathematica and Wolfram|One)

You'll Learn To

  • Build and train your own networks
  • Symbolically represent network layers without worrying about implementation details
  • Save and load trained network models
  • Use flexible specifications for training data and losses
  • Access prepackaged datasets