Fundamentals of Generative AI and Large Language Models: Theory and Practice

Übersicht

ABOUT THE COURSE:

This course provides a rigorous exploration of modern generative AI, covering latent-variable models such as Autoencoders, VAEs, GANs, and Diffusion Models that form the backbone of contemporary generative systems.Learners will develop a deep understanding of probabilistic modeling, adversarial training dynamics, denoising diffusion processes, and latent space manipulation for high-fidelity data generation.The sequence modeling module bridges classical NLP with modern architectures, introducing RNNs, LSTMs, and culminating in the Transformer framework that powers today’s Large Language Models.Emphasis is placed on mathematical foundations, architectural intuition, and practical implementation using widely adopted deep learning libraries.By the end of the course, learners will be equipped to analyze, design, and implement state-of-the-art generative models and LLM-based systems across diverse application domains.INTENDED AUDIENCE:

UG and PG students of all the AICTE affiliated institutionsPREREQUISITES:

Students must have completed introductory courses in Programming and Machine Learning/Deep Learning.Knowledge of Python and basic mathematical concepts is necessary to follow the hands-on exercises.INDUSTRY SUPPORT:

Generative AI and Large Language Models are now core technologies adopted across global and Indian industries, including software, healthcare, finance, retail, manufacturing, and creative design. Companies such as Google, Microsoft, Meta, Amazon, NVIDIA, IBM, OpenAI, and leading Indian organizations like TCS, Infosys, Wipro, LTIMindtree, Tech Mahindra, and Cognizant actively recruit professionals skilled in generative modeling and LLMs.

AI-driven startups such as HuggingFace, Stability AI, Rephrase.ai, Sarvam AI, and Qure.ai also value these competencies for developing foundation models, multimodal AI systems, and domain-specific generative applications. This course equips learners with the theoretical and practical expertise highly recognized and sought after across these industries.

Lehrplan

• Introduction to Generative AI
Overview of Generative Models
Applications in Industry
• Latent-Variable Models
Autoencoders
Variational Autoencoders (VAEs)
Generative Adversarial Networks (GANs)
Basics of Probabilistic Modeling
• Advanced Generative Models
Diffusion Models
Denoising Diffusion Processes
Latent Space Manipulation
• Sequence Modeling and Natural Language Processing
Classical NLP: RNNs and LSTMs
Introduction to Transformers
Large Language Models (LLMs)
• Mathematical Foundations
Probability and Statistics for Machine Learning
Linear Algebra and Calculus for Deep Learning
• Deep Learning Architecture and Intuition
Understanding Neural Network Architectures
Adversarial Training Dynamics
• Practical Implementation
Hands-on with Autoencoders and VAEs using PyTorch/TensorFlow
Building GANs using Deep Learning Libraries
Implementing Diffusion Models and Basic Transformers
• Contemporary Applications and Case Studies
High-fidelity Data Generation
Multimodal AI Systems
Domain-specific Generative AI Applications
• Course Project
Designing and Implementing a Generative Model
Real-world Application Development
• Conclusion and Future Trends in Generative AI
Evolving Architectures and Techniques
Ethical Considerations and Best Practices
• Evaluation
Quizzes and Assignments
Final Project Presentation and Report

---

Unterrichtet von

Prof. Sriram Ganapathy, Prof. Ashwini Kodipalli, Prof. Baishali Garai

---

Fachgebiete

Artificial Intelligence