Machine Learning and Applications

Daniel L.Silver

Machine Learning and Applications

•
Sub
-
area of Artificial Intelligence

•
Induction = development of predictive models
from examples

•
Neural networks, decision trees, genetic algorithms

•
Application to Data Mining

–
Medical diagnosis

–
Target marking, web mining



Daniel L.Silver

Machine Learning and Inductive Transfer

Environment


X

Training

Examples

Testing

Examples


(
x, f
(
x
))

Model of

Classifier

h

Inductive

Learning System

short
-
term memory

h
(
x
) ~
f
(
x
)

Domain

Knowledge

long
-
term memory

Retention &

Consolidation

Inductive

Bias

Selection

Knowledge

Transfer

Daniel L.Silver

Coronary Artery Disease Diagnosis

Results on California data (20 training examples) after
learning Cleveland & Hungary models

0.56

0.65

0.67

0.55

0.59

0.73

0.56

0.66

0.66

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Accuracy

Sensitivity

Specificity

Baseline

Task Rehearsal

Task Rehearsal

+ Relatedness

Stream Flow Rate Prediction

Stream flow rate prediction [Lisa Gaudette, 2006]


x
= weather data



f(x)

= flow rate

Daniel L.Silver

Image Transformation

Work withJude Abbey/Liangliang Tu, 2006
-
08

Daniel L.Silver

User Modeling

Intelligent Web Filters

Form Field Ordering

and Completion

Handheld Fashion

Consultant

Smart Email Client

Daniel L.Silver

User Identification

Key Stroke Biometrics

Smart Navigator

Handwriting ID

Eye
-
tracking Biometrics