Machine Learning and Applications

Preview:

Citation preview

MACHINE LEARNING AND APPLICATIONS Geeta Arora

Expert Session delivered during Workshop on Image Processing and Machine Learning for Pattern Recoginition on 11th July 2016 at University Institute of Engineering and Technology, Chandigarh

What is Machine Learning?

Hello World of Machine Learning

Types of Machine Learning

Why Machine Learning?

AGENDA

Libraries

Applications

Demo

Workflow

What is Machine Learning?

Hello World of Machine Learning

Types of Machine Learning

Why Machine Learning?

AGENDA

Libraries

Applications

Demo

Workflow

Machine Learning is not the future

Why Machine Learning?

Develop systems that can automatically adapt themselves to users

• Personalized news or emails

Discover new knowledge from large databases

Ability to mimic humans and replace monotonous tasks which require some intelligence

• Recognizing handwritten characters

Develop systems that are too difficult or expensive to

construct manually or knowledge tuned to a

specific task

Why the increased interest in Machine Learning? 1. Growing volumes and varieties of available data

2. Cheaper and more powerful computational processing

3. Better and inexpensive storage capacities

4. Open source revolution

What is Machine Learning?

Hello World of Machine Learning

Types of Machine Learning

Why Machine Learning?

AGENDA

Libraries

Applications

Demo

Workflow

What is Machine Learning?

Field of study that gives the computers the ability to learn without being explicitly programmed. (Arthur Sameul, 1959)

What is Machine Learning?

Hello World of Machine Learning

Types of Machine Learning

Why Machine Learning?

AGENDA

Libraries

Applications

Demo

Workflow

Problem –

Distinguish Apple from Orange

Write a function which takes and image file and outputs if it is an apple or not?

Start writing rules…

Rules start breaking ….

Lets solve with machine learning

Classifier

Apple

How do you solve the problem using Machine Learning?

Collect Training Data

Train Classifier Make

Predictions

Collect Training Data

Train Classifier Make

Predictions

Feature 1 Feature 2 Example

Example

The Code – Hello World of Machine Learning

from sklearn import tree features = [[140, "smooth"], [130, "smooth"], [150, "bumpy"], [170, "bumpy"]] labels = ["apples", "apples", "orange", "orange"] features = [[140, 1], [130, 1], [150, 0], [170, 0]] labels = [0, 0, 1, 1] clf = tree.DecisionTreeClassifier() clf.fit(features, labels) print(clf.predict([[160, 0]]))

clf.fit(features, labels) Under the hood…

What is Machine Learning?

Hello World of Machine Learning

Types of Machine Learning

Why Machine Learning?

AGENDA

Libraries

Applications

Demo

Workflow

Types of Machine Learning

Machine Learning

Supervised

Regression

Classification

Unsupervised Clustering

Reinforcement Recommendations

What is Machine Learning?

Hello World of Machine Learning

Types of Machine Learning

Why Machine Learning?

AGENDA

Libraries

Applications

Demo

Workflow

Machine Learning Workflow

What is Machine Learning?

Hello World of Machine Learning

Types of Machine Learning

Why Machine Learning?

AGENDA

Libraries s

Applications

Demo

Workflow

Machine Learning Libraries

What is Machine Learning?

Hello World of Machine Learning

Types of Machine Learning

Why Machine Learning?

AGENDA

Libraries

Applications

Demo

Workflow

Gmail: ensuring a spam-free inbox with Machine Learning

Google's Smart Email Reply

Google Photos uses machine learning to create customized albums

Google Keyboard

Google Speech Recognition

Apple Siri

Amazon Echo

https://play.google.com/store/apps/details?id=com.scigh.cricketworldcup2015

DEMO Scigh Cricket App

geeta.a.arora@gmail.com

Recommended