Generative AI Course

Basics of Python Programming
Installation and Setup:
Installing Python and setting up a development environment (IDEs
like PyCharm, VSCode, Jupyter Notebooks)

Syntax and Basic Constructs:

• Variables and data types (integers, floats, strings, booleans)
• Basic input and output
• Comments and documentation
• Control Structures

Conditional Statements:
 if, elif, else
 Loops:

• for, while
• Loop control statements (break, continue, pass)

Functions
Defining Functions:

• Parameters and return values

Scope and Lifetime:

• Local and global variables

Lambda Functions:

• Anonymous functions

• Introduction to Generative AI
• AI vs ML vs DL vs NLP vs Generative AI
• Generative AI principles
• What is the role of ML in Gen-AI
• Different ML techniques (Supervised, Unsupervised, Semisupervised & Reinforcement Learning)
• Applications in various domains
• Ethical considerations

• NLP essentials
  ● Basic NLP tasks
  ● Different text classification approaches
  ● Frequency based – Bag of words,TF-IDF, N-gram.
  ● Distribution Models – CBOW, Skipgram(Traditional approaches)and
  word2vec, Glove.
  ● Ensemble Methods (Random Forest, Gradient Boosting, AdaBoost) &
  Traditional Machine Learning Models – Naïve Bayes, Support Vector
  Machine (SVM), Decision Trees, Logistic Regression.
  ● Deep learning techniques – CNNs, RNNs, LSTMs, GRU and
  Transformers.

• Autoencodes
  ● VAE’s and applications
  ● GAN’s and it’s applications
  ● Different types of GAN’s and applications

• Different types of Language models
  ● Applications of Language models
  ● Transformers and its architecture
  ● BERT, RoBERTa, GPT variations
  ● Applications of transformer models

• What is Prompt Engineering
  ● What are the different principles of Prompt Engineering
  ● Types of Different Prompt Engineering Techniques
  ● How to Craft effective prompts to the LLMs
  ● Priming Prompt
  ● Prompt Decomposition

• Generative AI lifecycle
  ● What is RLHF
  ● LLM pre-training and scaling
  ● Different Fine-Tuning techniques

• What are word embeddings
  ● What is the use of word embeddings, where we can use it?
  ● Word Embeddings – Word2Vec, GloVe and FastText
  ● Contextual Embeddings – ELMo , BERT and GPT
  ● Sentence Embeddings – Doc2Vec, Infersent, Universal Sentence
  Encoder
  ● Subword Embeddings – BPE(Byte Pair Encoding), Sentence Piece
  ● Usecase of Embeddings.

• What is Chunking
  ● What is the use of chunking the document
  ● What are the traditional effective chunking techniques
  ● What are the problems and limitations with traditional chunking
  techniques?
  ● How to overcome the limitations of Traditional chunking
  ● Advanced Chunking Techniques:
  1. Character Splitting
  2. Recursive Character Splitting
  3. Document based Chunking
  4. Semantic Chunking
  5. Agentic Chunking

• What is RAG
  ● What are the main components of RAG
  ● High level architecture of RAG
  ● How to Build RAG using external data sources
  ● Advanced RAG

• What is Langchain
  ● What are the core concepts of Langchain
  ● Components of Langchain
  ● How to use Langchain agents

• LlamaIndex
  ● What are Vector Databases
  ● Why do we prefer Vector Databases over Traditional Databases
  ● Different Types of Vector Databases: OpenSource and Close Source
  ● OpenSource: Chroma DB, Weaviate,Faiss,Qdrant
  ● Close-Source Vector Databases:Pinecone,ArangoDB,Cloud-Based
  Solutions

• Supervised Finetuning
  ● Repurposing-Feature Extraction
  ● Advanced techniques in Supervised Finetuning -PEFT -LoRA, QLoRA

• Text based LLMs
  Automatic Evaluation: BULE Score, ROUGE Score, METEOR, BERT
  Score.
  Human Evaluation: Coherence, Factuality, Originality, Engagement

  Image based LLMs
  Automatic Evaluation: Pixel-level metrics, FID (Frechet Inception
  Distance), IS (Inception Score), Perceptual Quality Metrics,
  Diversity Metrics.
  Human Evaluation: Photorealism, Style, Creativity, Cohesiveness

  Audio generation LLMs
  Automatic Evaluation:  

  FAD (Frechet Audio Distance), 

  IS (Inception Score), 

  Perceptual Quality Metrics – PAQM, 

   PAQM – SNR (Signal-to-Noise Ratio), 

  PAQM – PESQ (Perceptual Evaluation of Speech Quality)
  Human Evaluation: 

  Perceptual Quality – PQ, 

   PQ- Naturalness,

   PQ-Fidelity, 

  PQ- Musicality, 

  Task Specific Evaluation.

  Video Generation LLMs
  Automatic Evaluation: FVD (Frechet Video Distance), Inception
  Score(IS), Perceptual Quality Metrics, Motion Based Metrics –
  Optical Flow Error, Content-Specific Metrics.
  Human Evaluation: Visual Quality, Temporal Coherence, Content
  Fidelit.

• Model Deployment and Management

•  Scalability and Performance Optimization

•  Security and Privacy

•  Monitoring and Logging

• Cost Optimization

• Model Interpretability and Explainability.

• Amazon Bedrock, Azure OpenAI

• ChatGPT, Gemini, Copilot