Module 1 — Machine Learning & Neural Network Basics
Published:
Deep Learning Tutorial Series
- Module 0 — Environment Setup (Beginner Guide)
- Module 1 — Machine Learning & Neural Network Basics
- Module 2 — Training & Learning Process
- Module 3 — Practical Deep Learning with PyTorch
- Module 4 — CNNs and Computer Vision
- Module 5 — Sequence Models + Modern Deep Learning Overview
This module introduces the core ideas behind machine learning and neural networks. The focus is on intuition rather than mathematics or code.
1.1 Learning Goals
By the end of this module, you should understand:
- What machine learning is
- Difference between traditional programming and ML
- Types of machine learning (supervised / unsupervised)
- What a model is
- Basic structure of a neural network
- How predictions are generated
1.2 What is Machine Learning?
Machine learning is a way for computers to learn patterns from data instead of being explicitly programmed with rules.
Traditional programming
Rules + Data → Output
Machine learning
Data + Output → Learn Rules (Model)
In ML, the system learns a function that maps inputs to outputs.
1.3 Types of Machine Learning
1) Supervised Learning
- Data includes inputs and correct answers (labels)
Example:
- Image → cat / dog
- Email → spam / not spam
Goal: learn mapping from input to output
2) Unsupervised Learning
- Data has no labels
Example:
- Clustering customers
- Finding patterns in data
Goal: discover hidden structure in data
3) Reinforcement Learning (overview only)
- Learning through rewards and feedback
- Example: game agents learning to play
1.4 What is a Model?
A model is a function that learns from data:
y = f(x)
Where:
- x = input
- y = output prediction
- f = learned mapping
Training means learning the best function f.
1.5 What is a Neural Network?
A neural network is a model made of layers of simple computational units called neurons.
Structure:
- Input layer
- Hidden layers
- Output layer
Each layer transforms the data step by step.
1.6 The Neuron (Core Idea)
A neuron performs:
output = activation(w1*x1 + w2*x2 + ... + b)
Where:
- x = inputs
- w = weights (learned importance)
- b = bias
- activation = non-linear function
The neuron learns which inputs matter most.
1.7 Activation Functions
Without activation functions, a neural network would only be a linear model.
ReLU
f(x) = max(0, x)
- Most commonly used
- Simple and effective
Sigmoid
- Outputs between 0 and 1
- Often used for binary classification
Tanh
- Outputs between -1 and 1
- Less commonly used in modern deep learning
1.8 Forward Pass (Prediction Process)
A forward pass is how a neural network makes predictions:
Input → Layers → Output
Steps:
- Input enters network
- Each layer applies weights + activation
- Final layer produces prediction
1.9 Simple Example
Predicting house price:
Inputs:
- Size
- Location
- Number of rooms
Output:
- Price
The model learns how each feature affects the final price.
1.10 What Does “Learning” Mean?
Learning means improving predictions by adjusting weights.
High-level process:
- Make prediction
- Compare with correct answer
- Measure error
- Adjust parameters
(Training details will come in Module 2)
1.11 Key Terms Summary
| Term | Meaning |
|---|---|
| Model | Function that maps input to output |
| Feature | Input variable |
| Label | Correct output |
| Training | Learning from data |
| Prediction | Model output |
1.12 Quick Check (No Coding)
Think about:
- What inputs would a weather prediction model need?
- What is the output?
- Is this supervised or unsupervised learning?
Acknowledgement
Part of the contents are generated by ChatGPT.
Return to the Main Page of Deep Learning and Machine Learning .