You are on page 1of 5

Prepared By:

2021-ME-108(SDP)
2021-ME-139(SDP)
Categories of Artificial
Neural Network
• The monolayer perceptron has a
single layer of neurons and is suitable
for simple binary classification tasks.
• The bi-layer MLP has two hidden
layers and can be used for more
complex tasks like image classification.
• The tri-layer MLP has three
hidden layers and is useful for even
more complex tasks like natural
language processing (NLP) tasks.
Purpose Of Hidden Layer

• Feature Extraction: Hidden layers identify and extract


important data features.
• Hierarchical Representation: They build complex data
representations layer by layer.
• Non-Linearity: Hidden layers enable ANNs to handle
non-linear relationships(e.g., sigmoid, ReLU).
• Increased Model Capacity: They enhance the
network's ability to tackle diverse and complex tasks
Effect Of Neurons On ANN Output
Neurons are fundamental computational units inspired by biological neurons. They are organized into layers and
connected to process information.

• Information Processing: Neurons process input data to generate meaningful


information.
• Feature Recognition: They detect patterns and features in the data.

• Non-Linearity: Neurons introduce non-linear transformations for modelling


complex relationships.

• Weight Adaptation: Neurons' weights adjust during training to improve accuracy.

• Final Prediction: Neurons in the output layer produce the network's ultimate
predictions.

You might also like