Back-Propagation Algorithm –
Addl. Issues
Neural Networks (Machine
Learning)
• Neurons
• accept information from multiple inputs,
• transmit information to other neurons.
• Complexity of Human Biological Neural Network
Supervised learning
• Definition
• How is it different from un-supervised learning?
• How to determine NN algorithm as supervised or un-supervised?
• The learning law
• Learning law is like heart of NN algorithm
• Target / Desired value
• Iterative process until convergence / target is achieved
3
Artificial Neural Networks – Learning
laws
4
Pattern Classification
5
Pattern Classification Task - Example
6
Pattern mapping tasks
7
Pattern mapping – An illustration
8
Forecast / Prediction problem using
Back-Propagation Algorithm
9
2-Dimensional Input layer of FFNN
10
Mesh Vs Layered Topology
Mesh Topology Layered Topology
Source: Source: [Link]
[Link] forward-neural-network
[Link]?TocPath=DigiMesh%20networks%7C
Machine Learning
Training Training
Labels
Training
Images
Image Learned
Training
Features model
Testing
Image Learned
Prediction
Features model
Test Image Slide credit: D. Hoiem and L.
Features
• Feature Vector
• Significance of Features
• Dimensionality issues
• How the patterns are represented
• With discriminatory features
• Forming feature vectors (FVs)
Ordering of Features is Important
Deep Learning
Imp. Characteristics of Deep Learning:
• No feature Extraction
• A large number of layers
Source: [Link]
Architecture of Convolutional Neural
Networks (CNNs)
Image Source:https
://[Link]/guides/convolutional-neural-networks/convolutional-neural-network-tutorial-basic-advan
ced
/
THANK YOU
Ethnocentrism
