Skip to main content

Deep Learning Foundations

Practicum AI

Practicum AI is a hands-on applied artificial intelligence (AI) curriculum intended for learners with limited coding and math background. To register for these courses, please fill out the registration form. You do not need to register for all courses and may instead register for only the courses that are most relevant to your work. You will need a SCINet account for this course. If you do not have a SCINet account, you may request one.

Course 4 – Deep Learning Foundations

Deep learning is the focus of modern AI. Models have many layers and millions, or now approaching a trillion, parameters! This course breaks things down and introduces you to a small AI model to provide a conceptual understanding of how AI models are built, trained, and deployed.

Prerequisites – Python for AI content: Basics of Python coding and Pandas
Objectives – Learners will be able to:

  1. Define neural networks, their basic components (neurons, perceptrons, bias, weights), and activation functions.
  2. Understand the significance of deep networks in neural computation.
  3. Utilize deep learning for image tasks, from classification to training with Keras’ MNIST package.
  4. Dive into TensorFlow and perceptron implementation.
  5. Grasp key concepts like gradient descent, optimizers, and the chain rule.
  6. Discuss the role of loss functions in network training.
  7. Explore applications in image classification and Natural Language Processing using pre-trained models.