Selecting the LSTM Last Timestep

MaxPool Stride and Downsampling

Weight Scale and Exploding Values

Data Leakage in Scaling

Learning Rate Scheduler Placement

Batch Normalization Formula

Weight Initialization Strategy

Cross-Entropy Loss Computation

Dropout Behavior During Evaluation

PyTorch Training Loop

Forward Pass Through a Layer

SGD Parameter Update

Gradient of a Squared Error

Sigmoid Activation Implementation

Configuring the Adam Optimizer

Feedforward Network Output Layer

Counting Conv Layer Parameters

Conv Layer Input Channels

Counting Iterations per Epoch

Normalizing Test Data Correctly

Weights, Biases, and Weighted Sums

The Artificial Neuron / Perceptron

What a Neural Network Is

Real-World Applications of Deep Learning

CNN vs RNN Architectures Overview

Regularization: Dropout and Weight Decay

Feature Learning vs Manual Feature Engineering

What Makes a Network 'Deep'

Epochs, Batches, and Iterations

Hierarchical Feature Learning

Purpose of Activation Functions

Epochs, Batches, Iterations

CNNs Overview

Learning Rate Effect

Why Non-Linearity Matters

Backpropagation Concept

Weight Initialization Importance

Perceptron Structure

Gradient Descent Training

Vanishing Gradients

Forward Propagation

Weights and Biases

Biological Neuron Inspiration

Network Layers

What Makes a Network Deep

Weighted Sum Computation

Common Activation Functions

Fully Connected Layers

RNNs Overview

Dropout Regularization

Learning Rate and Its Effect on Training

Backpropagation Basics

Gradient Descent for Optimization

Loss Functions for Neural Networks

Forward Propagation Basics

Neural Network Layers Explained

Why Deep Learning Needs Data and Compute

Recurrent Neural Networks (RNNs)

Convolutional Neural Networks (CNNs)

Artificial Neuron / Perceptron Structure

Loss / Cost Functions

Activation Functions: Purpose

Weights, Biases, and Weighted Sum

Vanishing and Exploding Gradients

Epochs, Batches, and Iterations

Input, Hidden, and Output Layers

Gradient Descent Intuition

Backpropagation

What Makes a Network 'Deep'

Learning Rate

Hyperparameters vs Learned Parameters

Common Activation Functions

Feature / Representation Learning

Regularization (Dropout, etc.)

Neural Networks and Biological Inspiration

Feedforward and Forward Propagation

Activation Functions: Sigmoid, ReLU, Tanh