Wolfram U

Neural Networks Basic Concepts

Estimated Time: 30 min

Course Level: Beginner

Summary

Learn the fundamentals of neural networks and even train your own convolutional network. Discover how Wolfram Language provides an integrated, high-level approach to interfacing with neural nets, enabling you to skip the low-level programming and focus on building sophisticated networks symbolically. This video also demonstrates how you can leverage neural nets for your own varied applications. 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 video starts at a beginner level with a review of basic concepts like tensors, layers and training and then goes on to illustrate how to train an entire network from scratch.

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

You'll Learn To

  • Grasp the fundamental ideas underlying neural networks
  • Build and train your own networks
  • Symbolically represent network layers without needing to worry about implementation details
  • Save and load trained network models
  • Use flexible specifications for training data and losses
  • Access prepackaged datasets