Fundamentals of Deep Learning and Development of Generative AI Models Training

Duration: Five Days

Overview

In this Generative AI course, students learn the fundamentals of deep learning and how to train generative AI models, starting with a review of core Python concepts if needed. You will also learn about the Anaconda computing environment, importing and manipulating data with Pandas, and exploratory data analysis with Pandas and Seaborn.

Objectives

• Understand the basics of machine learning (ML) and deep learning, including the different types of ML models, classification and regression, and neural networks.
• Develop a deep learning model construction for prediction, including preprocessing tabular datasets for deep learning workflows, data validation strategies, architecture modifications for managing overfitting, and regularization strategies.
• Apply trustworthy AI frameworks for this deep learning prediction context.
• Learn the fundamentals of generative AI, including generating new content versus analyzing existing content, example use cases, and the ethics of generative AI.
• Implement sequential generation with recurrent neural networks (RNNs) and variational autoencoders (VAEs).
• Build a generative adversarial network (GAN).
• Understand transformer architectures, including the problems with recurrent architectures, attention-based architectures, positional encoding, and the Transformer model.
• Gain an overview of current popular large language models (LLMs), such as ChatGPT, DALL-E 2, and Bing AI.
• Learn about medium-sized LLMs that can be run in your environment, such as Stanford Alpaca and Facebook Llama.
• Explore transfer learning with your data in these contexts.

Prerequisites

Learners should have prior experience developing Deep Learning models, including architectures such as feed-forward artificial Neural Networks, recurrent and convolutional.