Advanced Artificial intelligence Course

Lesson 1 • Setting Up the AI Development Environment

Configures Python-based toolchains, version control, and GPU-enabled compute. Ensures students can run experiments from chapter two onward.

Lesson 2 • AI Problem Types and Taxonomies

Categorises problems as search, optimisation, classification, or generation. Enables students to map real tasks to the correct AI approach.

Lesson 3 • Mathematics for AI Practitioners

Reviews linear algebra, probability, and calculus essentials needed for AI algorithms. Bridges mathematical theory to practical model building.

Lesson 4 • History and Evolution of AI

Traces AI from symbolic logic to modern deep learning, highlighting paradigm shifts. Provides context for understanding why current techniques dominate.

Lesson 1 • Supervised Learning Fundamentals

Covers regression and classification algorithms, loss functions, and training loops. Forms the backbone for all predictive modelling work ahead.

Lesson 2 • Unsupervised Learning Techniques

Introduces clustering, dimensionality reduction, and density estimation without labels. Prepares students for exploratory data analysis and representation learning.

Lesson 3 • Model Evaluation and Selection

Teaches metrics, cross-validation, and bias-variance tradeoff analysis. Enables rigorous comparison of competing models before deployment.

Lesson 4 • Regularisation and Hyperparameter Tuning

Applies L1, L2, dropout, and search strategies to prevent overfitting. Directly improves model generalisation on unseen data.

Lesson 1 • Debugging and Diagnosing Deep Models

Identifies common failure modes: vanishing gradients, dead neurons, and data leakage. Builds systematic troubleshooting skills applicable to any architecture.

Lesson 2 • Training Deep Networks at Scale

Addresses batch normalisation, learning rate schedules, and distributed training. Equips students to train large models efficiently on real hardware.

Lesson 3 • Neural Network Fundamentals

Explains neurons, activation functions, forward pass, and backpropagation. Provides the mechanistic understanding needed to debug and extend any network.

Lesson 4 • Recurrent and Sequence Models

Teaches RNNs, LSTMs, and GRUs for temporal and sequential data modelling. Bridges to transformer-based sequence models introduced later.

Lesson 1 • Attention Mechanisms Explained

Derives scaled dot-product attention and multi-head attention from first principles. Provides the theoretical core of the transformer architecture.

Lesson 2 • Fine-Tuning Large Language Models

Applies supervised fine-tuning, prompt engineering, and parameter-efficient methods. Produces task-specific models without full retraining from scratch.

Lesson 3 • NLP Evaluation and Benchmarking

Uses BLEU, ROUGE, perplexity, and human evaluation to assess NLP models. Connects model quality metrics to real-world task performance.

Lesson 4 • Transformer Architecture Deep Dive

Dissects encoder, decoder, and encoder-decoder transformer variants. Enables students to select the right variant for each NLP task type.

Lesson 1 • Semantic and Instance Segmentation

Teaches pixel-level classification and instance mask prediction for scene understanding. Builds skills for autonomous systems and medical imaging applications.

Lesson 2 • Object Detection Architectures

Compares single-stage and two-stage detectors for real-time and high-accuracy use cases. Enables selection and adaptation of detection models for production.

Lesson 3 • Multimodal AI Systems

Integrates vision and language through contrastive learning and cross-modal attention. Enables building systems that reason across image and text simultaneously.

Lesson 4 • Generative Models for Images

Covers GANs, VAEs, and diffusion models for image synthesis and editing. Prepares students for creative AI and data augmentation workflows.

Lesson 1 • Actor-Critic Architectures

Combines value estimation and policy learning in a unified neural framework. Improves sample efficiency over pure policy gradient approaches.

Lesson 2 • Policy Gradient Methods

Derives REINFORCE and variance reduction techniques for direct policy optimisation. Extends the RL foundations from chapter two to continuous action spaces.

Lesson 3 • Reward Function Design

Addresses sparse rewards, reward shaping, and inverse RL from expert demonstrations. Directly impacts agent behaviour quality in real deployments.

Lesson 4 • RL in Simulated Environments

Uses physics simulators and game engines to train and benchmark RL agents safely. Prepares students for sim-to-real transfer challenges.

Lesson 1 • CI/CD for Machine Learning

Integrates automated testing, validation gates, and deployment pipelines for ML code. Applies software engineering discipline to AI system delivery.

Lesson 2 • Data Pipeline Engineering

Designs scalable ingestion, validation, and feature engineering workflows. Ensures data quality and reproducibility across all downstream model training.

Lesson 3 • Model Serving and Inference Optimisation

Deploys models via REST APIs, batch jobs, and edge devices with latency targets. Covers quantisation and pruning to meet real-time performance requirements.

Lesson 4 • Monitoring and Observability in Production

Detects data drift, model degradation, and system failures in live deployments. Closes the feedback loop between production signals and retraining triggers.

Lesson 1 • Explainability and Interpretability

Applies SHAP, LIME, and attention visualisation to explain model decisions. Enables compliance with transparency requirements and builds user trust.

Lesson 2 • Bias and Fairness in AI

Identifies sources of algorithmic bias and applies fairness metrics to audit models. Connects technical fixes to organisational accountability practices.

Lesson 3 • AI Safety and Robustness

Tests models against adversarial attacks, distribution shift, and failure modes. Builds defences that maintain reliable behaviour under real-world conditions.

Lesson 4 • AI Governance and Regulatory Frameworks

Maps AI risk tiers, documentation requirements, and accountability structures to governance policies. Prepares students to lead responsible AI programmes in organisations.