CEP IIT Delhi Certificate Programme in Generative AI - Batch 03

Module 1: Mathematical Foundations for ML

• Linear Algebra: Vector and Matrix Dot Product, Matrix Vector Multiplication, Matrix Decomposition (SVD)s
• Probability Theory: Random Variables, Bayes Theorem, Conditional Probability
• Optimisation: Gradient Descent, First/Second Order Condition, Convex Optimisation, KL-Divergence

Learning Outcomes:

• Building mathematical foundations, an essential prerequisite for the course
• Understanding the fundamentals of Probability Theory in Machine Learning
• Concepts of Optimisation and its application in Machine Learning algorithms

Module 2: Machine Learning

• Intro to ML: Linear Regression, Logistic Regression
• Optimisation Continued, SVM, Decision Tree, Ensemble Methods
• Unsupervised Learning: Clustering, Dimensionality Reduction (PCA, LDA, t-SNE)
• Artificial Neural Networks: Perceptron, Multilayer Network, Backpropagation

Learning Outcomes:

• Fundamental ML concepts, understanding regression methods
• Learning supervised ML methods
• Introduction to unsupervised methods
• Understanding the basics of neural networks and their training

Module 3: Natural Language Processing (NLP)

• Basic Text Processing (NLTK, spaCy), Morphology, Stemming, Edit Distance
• Language Modelling: N-gram Modelling, Smoothing Techniques, Perplexity
• POS Tagging: Sequential Learning, HMM, Viterbi Algorithm
• Parsing: Constituency vs Dependency Parsing, CKY Algorithm, CFG, PCFG
• Text Classification: Naive Bayes Algorithm, Lexical Similarity (Word Embeddings, TF-IDF, Word2Vec, GloVe)

Learning Outcomes:

• Use of NLP tools for text processing
• Learning different language modelling approaches
• Introduction to sequence learning for NLP
• Learning techniques for syntactic analysis in NLP
• Understanding classification in NLP using traditional and modern approaches

Module 4: Generative AI for Text

• Neural Language Models (CNN, RNN, LSTM, GRU, Seq2Seq)
• Attention Mechanism: Self-Attention, Transformer Architecture
• Pre-trained Models: BERT, GPT, T5
• Fine-tuning Strategies: Task-specific Fine-tuning, Instruction Fine-tuning, Preference Tuning (RLHF, PPO)
• Prompting Strategies: In-context Learning, Chain of Thought, Knowledge Probing, Text Generation
• Augmented LLMs: Retrieval-Augmented Generation (RAG), Tool Augmented LLM

Learning Outcomes:

• Introduction to neural models and attention mechanisms for text generation
• Understanding self-attention and Transformer architectures for language models
• Architecture and training of different pre-trained large language models
• Learning various fine-tuning approaches to improve model performance
• Techniques for optimising LLM performance through effective prompting
• Leveraging RAG and tool augmentation to improve LLM efficiency and reasoning capabilities

Module 5: Generative AI for Vision

• Vision Language Models (VLM)

Learning Outcomes:

• Understanding how VLMs enable combined text and image generation for multimodal applications

Module 6: Responsible AI

• Misinformation, Bias, Toxicity, Security and Fairness

Learning Outcomes:

• Learning strategies to mitigate bias, toxicity, and hallucinations in AI outputs

Module 7: Python Modules

• Python Modules
• Fundamentals of Python Programming
• Data structures and functional programming in python
• Python Libraries for Data science - Numpy and Pandas
• Python Libraries for Data science - matplotlib and seaborn
• Python code demos
• Case Study

Learning Outcomes

• Apply Python fundamentals
• Use data structures for problem-solving
• Analyze data using NumPy & Pandas
• Create visualizations

Module 8: EDA Modules

• Data Extraction
• SQL Primer
• Introduction to Data Wrangling
• Introduction to EDA
• EDA & Data Visualization with Tableau
• EDA code files
• Case study

Learning Outcomes

• Extract and query data using SQL
• Clean and analyze datasets
• Perform EDA
• Create data visualizations

Module 9: Agentic AI

• AI Agents and Agentic AI Systems
• Multi-Agent Systems and Agent Orchestration
• Conversational AI and Chatbot Development

Learning Outcomes

• Understand AI agents and systems
• Explore multi-agent workflows
• Work with basic chatbots

Module 10: Multimodal GenAI

• Multimodal Generative AI
• Diffusion Models and Foundation Models
• Audio and Speech Generation

Learning Outcomes

• Understand multimodal AI concepts
• Learn diffusion & foundation models (basic)
• Identify audio/speech AI applications

Assignments/Case Studies/Projects

• Implement a fully connected neural network using PyTorch or TensorFlow for some task. Train the model and visualise its accuracy and loss over epochs.
• Fine-tune a pre-trained transformer model on a text classification task. Evaluate the model on a test set and report its accuracy, precision, and recall.
• Implement a simplified version of the transformer architecture focusing on self-attention and positional encoding. Train the model on a small dataset for translation or another task. Compare the results with a pre-trained transformer model.
• Implement PEFT on a large LLM for a specific task. Compare the training time and resource usage between standard fine-tuning and PEFT.
• Create a simple reward model using human-labelled data for a text generation task (e.g., summarisation or sentiment correction). Implement a reinforcement learning loop to improve the LLM's responses based on the reward model.
• Compare the performance of different open-source Large Language Models (LLMs) on reasoning tasks using various prompting techniques—zero-shot, few-shot, chain-of-thought, self-consistency prompting.

Tutorials

• Compute matrix operations using NumPy.
• Preprocess a text dataset and generate embeddings using spaCy or Hugging Face.
• Build a simple neural network using TensorFlow/PyTorch.
• Train the model on basic NLP tasks like text classification problem.
• Implement a simplified self-attention mechanism.
• Use a pre-trained transformer models for various NLP tasks like translation or summarization.
• Fine-tune a GPT or BERT model on a custom dataset for various NLP problems
• Use Hugging Face's pre-trained transformer model to generate text.
• Apply PEFT to a large model for fine-tuning on a small dataset.
• Apply quantization and pruning on a pre-trained model to optimize for lower resource usage.
• Implement a RAG model using Hugging Face transformers and document retrieval from a custom corpus.
• Use Hugging Face multilingual models for problems in cross-lingual scenarios.