Machine Learning – Fundamentals and Applications

Neural Networks Demystified - Part 1: Data and Architecture – An introduction to neural networks, covering basic structure and how data flows through them.
Neural Networks Demystified - Part 2: Forward Propagation – Explains forward propagation with intuitive visuals.
Neural Networks Demystified - Part 3: Gradient Descent – Introduction to gradient descent for training neural networks.
Neural Networks Demystified - Part 4: Backpropagation – Detailed look at backpropagation.
Neural Networks Demystified - Part 5: Numerical Gradient Checking – Validating gradients using numerical methods.
Neural Networks Demystified - Part 6: Training – Training process explained.
Neural Networks Demystified - Part 7: Overfitting and Regularization – Handling overfitting in neural networks.

Deep Learning Specialization

Deep Learning Specialization - Intro – Overview of deep learning concepts.
Deep Learning Specialization - Neural Networks Basics – Fundamentals of neural networks.
Deep Learning Specialization - Improving Neural Networks – Hyperparameter tuning and optimization.
Deep Learning Specialization - Structuring ML Projects – How to organize ML projects.
Deep Learning Specialization - CNNs – Convolutional Neural Networks explained.
Deep Learning Specialization - Sequence Models – Understanding sequence models and RNNs.

Deep Learning – Intuitions

Why Deep Learning Works Unreasonably Well – A geometric and intuitive explanation of why deep learning performs so effectively.

AI and Society

How DeepSeek Rewrote the Transformer (MLA) – Explains the Multi-Head Latent Attention (MLA) technique introduced by DeepSeek, showcasing how it reengineers the Transformer architecture.

Visualizing Transformers

I Visualised Attention in Transformers – A simple and intuitive visual explanation of why attention is needed in Transformer models, illustrating what it actually does.
Visualizing transformers and attention | Talk for TNG Big Tech Day ‘24 – A comprehensive talk that dives into how Transformers work, including the attention mechanism, tokenization, parallelization, and practical applications across domains.