Basics of Generative AI and its evolution. Generative vs. Discriminative Models. Foundations of Deep Learning: Neural Networks, Backpropagation, and Activation Functions. Introduction to Key Architectures: Diffusion Models, Transformers, and Large Language Model (LLM) Architecture Overview. Critical Examination of Ethics, Risks, and Hallucination in AI Systems.
Language Model Families: GPT-series, LLaMA, Gemini. Vision Foundation Models: Stable Diffusion, Midjourney, DALL·E. Core Preprocessing: Tokenization and Embeddings. Introduction to Fine-tuning and Efficient Adaptation Methods: LoRA (Low-Rank Adaptation) and QLoRA.
Generative Painting and Illustration Models (e.g., Midjourney, Stable Diffusion for art). Music Generation Models: MuseNet, MusicLM. Script and Play Generation using LLMs. Introduction to Multimodal AI Models (e.g., GPT-4V). Overview of Creative AI Applications in various industries
Types of Prompts: Zero-Shot, Few-Shot, Chain-of-Thought (CoT). System Prompts and Role Prompting for controlling model behavior. Techniques for Prompt Optimization and Evaluation. Safety, Bias Mitigation, and Alignment Prompting
Concept of Autonomous AI Agents. Core Components: Tool Use, Memory, and Planning. Introduction to Multi-Agent Systems. Agent Frameworks: LangChain Agents, AutoGen. Real-world Use Cases and Applications of Agentic AI.
Reference Book:
1 Ethan Mollick, "Co-Intelligence: Living and Working with AI", Penguin Life, 2024 COURSE OUTCOMES At the end of the course student should be able to:
Text Book:
1 Numa Dhamani & Maggie Engler, Introduction to Generative AI, Manning / Simon & Schuster, 2024 2 Dale Markowitz, Ari Bornstein, Practical Generative AI with Transformers: Text, Image, Audio & Multimodal Applications, O’Reilly Media, 2024.