KBCNMU · Third Year B.E. Artificial Intelligence & Data Science · NEP 2020
Introduction to Generative AI
AIDS-501 · Semester V · 3 credits · ISE 40 + ESE 60
Unit 1: Fundamentals of Generative AI
9 hoursDefinition and scope of generative AI, generative versus discriminative AI, kinds of output — text, images, video, music, code and voice — and real-world uses in banking, healthcare, manufacturing and retail.
Examined in the 60-mark end-semester paper
- Discriminative versus generative models
- Applications of generative AI
- Generative AI in healthcare, education and media
Unit 2: Generative AI Models — Landscape and Types
9 hoursText, image and code generative models, closed versus open-source models; GANs, variational autoencoders, transformers and attention in detail, LSTMs, and how the models compare.
Examined in the 60-mark end-semester paper
- Types of generative models: autoregressive, GANs, VAEs and diffusion
- Generative adversarial networks
- Autoencoders and VAEs
- Transformers and attention
- LSTM
Unit 3: Large Language Models, GPT and BERT
9 hoursWhat LLMs are, foundation models, pre-training, fine-tuning and RLHF; autoregressive generation and token-based inference; hallucination and limits; GPT and its variants, ChatGPT's architecture and capabilities; BERT's architecture, pre-training and fine-tuning.
Examined in the 60-mark end-semester paper
- Large language models: GPT, BERT, T5 and LLaMA
- Fine-tuning pretrained and masked language models
- Reinforcement learning from human feedback
- Hallucination, bias and deepfakes
Unit 4: Prompt Engineering and Generative AI Programming
9 hoursPrompt components and structure, zero-shot and few-shot prompting, design strategies and common patterns; programming LLM-based applications; retrieval-augmented generation and its uses.
Examined in the 60-mark end-semester paper
- Why prompt engineering matters
- Zero-shot, one-shot and few-shot prompting
- Prompt patterns
- Building with LangChain
- Retrieval-augmented generation
Unit 5: Agents, Enterprise Applications and Responsible AI
9 hoursAutonomous and agent-based generative AI systems; business impact, enterprise adoption and its challenges; responsible-AI principles — ethics, bias, safety and transparency — governance, compliance and future trends.
Examined in the 60-mark end-semester paper
- Autonomous agents and the future of generative AI
- Enterprise adoption of generative AI
- Alignment, safety and interpretability
- Copyright and legal frameworks for generative AI
Unit contents are summarised from the official KBCNMU syllabus . AiCanCode.org is not affiliated with Kavayitri Bahinabai Chaudhari North Maharashtra University (KBCNMU).