Download presentation
Presentation is loading. Please wait.
Published byἈριστομάχη Λιακόπουλος Modified over 6 years ago
2
Motivation Computers are good at some things… Calculating
Sorting lists Manipulating data i.e. anything involving a well-defined algorithm (involving “calculation”)
3
Motivation Computers are bad at some things… Recognizing faces
Noticing patterns Making decisions i.e. things that humans do all the time!
4
Examples Image/sound recognition Faces/voices Text Problems/anomalies
5
Examples Decision-making Loan application Path navigation
Path optimization Classification Sorting Who wrote this text?
6
Examples Prediction Stock market Football Purchasing habits/desires
10
Basic Idea of Neural Networks
Since humans can easily solve certain problems that computers have a hard time solving… Why not try to mimic the brain’s behavior with a computer program?
11
Basic Idea of Neural Networks
Humans: Observe input and/or output Numerous times Learn often by trial-and-error Extend observed cases to new observations
12
Basic Idea of Neural Networks
Computer model: Function to turn input into output Learn from observed data Brain = highly interconnected network of simple elements Learning = adjusting (strength of) connections
13
InputOutput: Examples
Image/sound recognition Faces/voices Text Problems/anomalies
14
InputOutput: Examples
Decision-making Loan application Path navigation Path optimization Classification Sorting Who wrote this text?
15
InputOutput: Examples
Prediction Stock market Football Purchasing habits/desires
16
Learning Supervised Reinforcement Unsupervised/associative
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.