Introduction to Neural Networks with Python

Neural networks are at the core of artificial intelligence and machine learning. These powerful models, inspired by the human brain, have the ability to learn from data and make predictions or decisions. In this post, we will provide a beginner-friendly introduction to neural networks, focusing on their basic architecture and implementation using Python. By the end, you will have a solid understanding of how neural networks work and be able to build one from scratch.

What are Neural Networks?

Neural networks are machine learning models designed to recognize patterns and relationships in data. They consist of interconnected nodes, called artificial neurons or “units,” organized into layers. Each neuron processes input from the previous layer and produces an output signal, which is then passed to the next layer. Neural networks learn from data through a training process where they adjust their internal parameters to make accurate predictions.

Basic Architecture of a Neural Network

A typical neural network consists of an input layer, one or more hidden layers, and an output layer. The input layer receives raw input data, while the output layer produces final predictions. The hidden layers transform the input data to enable accurate predictions.

Activation Functions

Activation functions introduce non-linearity into neural networks, enabling them to learn complex relationships in the data. Let’s look at some popular activation functions

Sigmoid Function

import numpy as np

def sigmoid(x):
    return 1 / (1 + np.exp(-x))

Rectified Linear Unit (ReLU)

def relu(x):
    return np.maximum(0, x)

Hyperbolic Tangent (tanh)

def tanh(x):
    return np.tanh(x)

Training a Neural Network

Training a neural network involves adjusting its parameters (weights and biases) iteratively to minimize the difference between predicted and actual outputs. This optimization is achieved through backpropagation, which calculates gradients of the loss function with respect to the network’s parameters. Let’s consider a simple example of training a neural network to solve a binary classification problem

# Assuming we have input data X and corresponding labels y
# Initialize the weights and biases
weights = np.random.randn(input_size, hidden_size)
biases = np.zeros(hidden_size)

# Forward pass
hidden_layer_output = relu(, weights) + biases)

# Calculate the loss (e.g., using cross-entropy)
loss = -np.mean(y * np.log(hidden_layer_output) + (1 - y) * np.log(1 - hidden_layer_output))

# Backpropagation
d_loss = (hidden_layer_output - y) / len(y)
d_hidden =, weights.T)
d_weights =, d_hidden * (hidden_layer_output * (1 - hidden_layer_output)))
d_biases = np.sum(d_hidden * (hidden_layer_output * (1 - hidden_layer_output)), axis=0)

# Update the parameters
weights -= learning_rate * d_weights
biases -= learning_rate * d_biases

Implementing a Neural Network in Python

Now, let’s implement a simple neural network using the popular deep learning library, TensorFlow. We will build a network to solve a regression problem

import tensorflow as tf

# Define the model architecture
model = tf.keras.models.Sequential([
    tf.keras.layers.Dense(32, activation='relu', input_shape=(input_size,)),
    tf.keras.layers.Dense(1)  # Output layer

# Compile the model
model.compile(optimizer='adam', loss='mse')

# Train the model, y_train, epochs=10, batch_size=32)

# Make predictions
predictions = model.predict(X_test)


Neural networks are powerful tools in the realm of machine learning. In this post, we provided an introductory guide to neural networks, covering their basic architecture, activation functions, training process, and implementation in Python. Armed with this knowledge and the code examples provided, you are ready to dive deeper into the world of neural networks. Experiment with different architectures, activation functions, and datasets to unlock the full potential of neural networks in solving real-world problems.

Additionally, if you have any further questions or want to explore neural networks in more depth, feel free to leave a comment below. If you prefer personalized 1-to-1 training sessions or have specific topics you’d like to delve into, you can also contact me for tailored training sessions.


5 1 vote
Article Rating
Notify of
Inline Feedbacks
View all comments
Would love your thoughts, please comment.x