SlideShare a Scribd company logo
MACHINE LEARNING
PRINCIPLES & ALGORITHMS
OUTLINE
• What is Machine Learning?
• Applications in Machine Learning
• (The Machine Learning) Model
• Machine Learning Models in Action
• Training Data
• Model / Data Considerations
• Models
• DecisionTree
• Random Forest
• Clustering
• Linear Models
• SupportVector Machines (SVM)
• Artificial Neural Networks
• Deep Learning (CNN)
• Reinforcement Learning
WHAT IS MACHINE LEARNING?
“Field of study that gives computers the ability to
learn without being explicitly programmed.”
- Arthur Samuel
A computer program is said to learn from experience E with respect to
some taskT and some performance measure P, if its performance onT,
as measured by P, improves with experience E.
-Tom Mitchell1959 1998
WHAT IS MACHINE LEARNING?
APPLICATIONS INTHE MODERN WORLDAPPLICATIONS INTHE MODERN WORLD
Optical Character
Recognition
Recommendation
Engines
Facial Recognition
Autonomous
Vehicles
Personal Assistants /
Chat Bots
MODEL
A REPRESENTATION OF A REALWORLD PROCESS
Water Cycle
MODEL
A REPRESENTATION OF A REALWORLD PROCESS
Water Cycle
Evolution
MODEL
A REPRESENTATION OF A REALWORLD PROCESS
Water Cycle
Evolution
Neuron-McCulloch & Pitts Model, 1943
MACHINE LEARNING MODELS IN ACTION
Untrained Model(Old)
DATA
Trained Model New
Data
Info?
Prediction?
Decision?
Expert
Knowledge
TRAINING DATA
Feature =Variable = Predictor Objective Measurement
Height (in) Weight (lb) Color Claws retract Class
11.2 10.1 black yes cat
23.1 45.2 black/white no dog
13.0 20.1 black/white yes cat
9.7 7.2 white yes cat
… … … … …
TRAINING DATA
Feature =Variable = Predictor Objective Measurement
Height (in) Weight (lb) Color Claws retract Class
11.2 10.1 black yes cat
23.1 45.2 black/white no dog
13.0 20.1 black/white yes cat
9.7 7.2 white yes cat
… … … … …
TRAINING DATA
Feature =Variable = Predictor Objective Measurement
Height (in) Weight (lb) Color Claws retract Class
11.2 10.1 black yes
23.1 45.2 black/white no dog
13.0 20.1 black/white yes
9.7 7.2 white yes cat
… … … … …
TESTING DATA (NO PEEKING!)
Training and testing sets must
ALWAYS be disjoint
• Cross-validation
• Leave-one-out
• OOB (Out-of-bag for
ensembles)
MODEL/DATA CONSIDERATIONS
(RELEVANT TO MODEL SELECTION)
Each model can/cannot handle certain data characteristics / analysis needs
• Supervised vs. Unsupervised data?
• Class Imbalance (200 cats vs. 3 dogs)
• 2-class vs. Multiclass (say 200 cats, 146 dogs, 25 sugar gliders, 5 platypuses)
• Scale issues (see Distance-based Clustering; Normalization / Standardization)
• FeatureType (Categorical, Continuous, etc)
• Dimensionality (# of features / measurements)
• Cost Sensitivity (Miss / False Alarm – can the model adjust?)
• Propensity to Overtrain (fitting to noise – see Bias vs.Variance)?
• Need to estimate uncertainty?
• Ability to adapt to changing conditions (parameters)?
• Robustness to sparse data (parameter estimation)?
DECISIONTREE
1) At each node, a question is asked
about a specific feature
2) The answer directs data left/right
3) Decision trees must be pruned to
prevent overtraining
RANDOM FOREST
Random Forest is an ENSEMBLE of DecisionTrees
RANDOM FOREST
RANDOM FOREST
Random Forest is an ENSEMBLE of DecisionTrees
Node Splits (Training)
• Bagging (resampled data for each
tree)
• “Best” univariate split on random
subspace (subset of all features)
• Gini Impurity
• Leaf nodes are class homogeneousLeo Breiman
RANDOM FOREST
Random Forest is an ENSEMBLE of DecisionTrees
Leo Breiman
Classification
1) Samples propagate through
each tree
2) Tree “votes” for a class
based on leaf node
3) Final decision based on class
conditional probability
CLUSTERING
KeyVariants
K-means: point-to-cluster mean distance
CLUSTERING
KeyVariants
K-means: point-to-cluster mean distance
Mean-Shift: hill-climbing to max density
CLUSTERING
KeyVariants
K-means: point-to-cluster mean distance
Mean-Shift: hill-climbing to max density
DBSCAN: epsilon neighborhood
CLUSTERING
KeyVariants
K-means: point-to-cluster mean distance
Mean-Shift: hill-climbing to max density
DBSCAN: epsilon neighborhood
Gaussian Mixture Models: Gaussian assumption
CLUSTERING
KeyVariants
K-means: point-to-cluster mean distance
Mean-Shift: hill-climbing to max density
Hierarchical Clustering
DBSCAN: epsilon neighborhood
Gaussian Mixture Models: Gaussian assumption
LINEAR MODELS
Linear Discriminant Analysis Simple Linear Regression
(Ronald)
Fisher’s LDA
SUPPORTVECTOR MACHINES (SVM)
Maps linearly nonseparable data to a higher dimension
Kernel trick
makes this mapping more
efficient
Also: sub-gradient descent, coordinate descent
SUPPORTVECTOR MACHINES (SVM)
Support vectors in the feature space used for classification
Support vectors are
determined by the
most difficult points to
classify…
ARTIFICIAL NEURAL NETWORKS
Recall the original model of the
neuron…
ARTIFICIAL NEURAL NETWORKS
Input Layer
Hidden Layers
Output Layer
Feedforward (forward processing)
• Each arrow represents a weight
• Hidden & output nodes “process”
input values/weights
Backpropagation (of errors)
• Allow specification of desired
output
• Minimize loss function
w11
w12
w41
f31
f41
f51
ARTIFICIAL NEURAL NETWORKS
Input Layer
Hidden Layers
Output Layer
w11
w12
w41
f31
f41
f51
DEEP LEARNING (CONVOLUTIONAL NN)
From the Latin convolvere,“to convolve” means to roll together
We convolve an image with multiple kernels (filters) at each layer
DEEP LEARNING (CONVOLUTIONAL NN)
Each layer of the network learns different features of the image
DEEP LEARNING (CONVOLUTIONAL NN)
REINFORCEMENT LEARNING
A reward-driven approach for a
machine to “self-learn”
• At each step, the agent takes an
action based on environment state
• The agent receives a reward based
upon the new state (post-action)
• The agent’s goal is to maximize his
reward
REINFORCEMENT LEARNING
Donald Michie creates MENACE, 1963
(Machine Educable Noughts And
Crosses Engine)
MENACE learned to play TicTacToe
using stacks of matchboxes
REINFORCEMENT LEARNING
Q(uality)-Learning – values-based; environment may be unknown
REINFORCEMENT LEARNING
Google’s
DeepMind AI
learns to walk
QUESTIONS / DISCUSSION

More Related Content

Similar to Carolina AI Meetup Nov 2018

Machine Duping 101: Pwning Deep Learning Systems
Machine Duping 101: Pwning Deep Learning SystemsMachine Duping 101: Pwning Deep Learning Systems
Machine Duping 101: Pwning Deep Learning Systems
Clarence Chio
 
Machine Learning
Machine LearningMachine Learning
Machine Learning
Arthur Breitman
 
Deep learning - a primer
Deep learning - a primerDeep learning - a primer
Deep learning - a primer
Uwe Friedrichsen
 
Deep learning - a primer
Deep learning - a primerDeep learning - a primer
Deep learning - a primer
Shirin Elsinghorst
 
AI and Deep Learning
AI and Deep Learning AI and Deep Learning
AI and Deep Learning
Subrat Panda, PhD
 
Machine Learning : why we should know and how it works
Machine Learning : why we should know and how it worksMachine Learning : why we should know and how it works
Machine Learning : why we should know and how it works
Kevin Lee
 
Machine Learning workshop by GDSC Amity University Chhattisgarh
Machine Learning workshop by GDSC Amity University ChhattisgarhMachine Learning workshop by GDSC Amity University Chhattisgarh
Machine Learning workshop by GDSC Amity University Chhattisgarh
Poorabpatel
 
What is Machine Learning
What is Machine LearningWhat is Machine Learning
What is Machine Learning
Bhaskara Reddy Sannapureddy
 
Intro to Machine Learning by Microsoft Ventures
Intro to Machine Learning by Microsoft VenturesIntro to Machine Learning by Microsoft Ventures
Intro to Machine Learning by Microsoft Ventures
microsoftventures
 
Echelon Asia Summit 2017 Startup Academy Workshop
Echelon Asia Summit 2017 Startup Academy WorkshopEchelon Asia Summit 2017 Startup Academy Workshop
Echelon Asia Summit 2017 Startup Academy Workshop
Garrett Teoh Hor Keong
 
Machine learning with R
Machine learning with RMachine learning with R
Machine learning with R
Maarten Smeets
 
Machine learning and linear regression programming
Machine learning and linear regression programmingMachine learning and linear regression programming
Machine learning and linear regression programming
Soumya Mukherjee
 
DeepLearning
DeepLearningDeepLearning
DeepLearning
ShahzadAsgharArain
 
Getting Started with Keras and TensorFlow - StampedeCon AI Summit 2017
Getting Started with Keras and TensorFlow - StampedeCon AI Summit 2017Getting Started with Keras and TensorFlow - StampedeCon AI Summit 2017
Getting Started with Keras and TensorFlow - StampedeCon AI Summit 2017
StampedeCon
 
The Power of Auto ML and How Does it Work
The Power of Auto ML and How Does it WorkThe Power of Auto ML and How Does it Work
The Power of Auto ML and How Does it Work
Ivo Andreev
 
Predicting Moscow Real Estate Prices with Azure Machine Learning
Predicting Moscow Real Estate Prices with Azure Machine LearningPredicting Moscow Real Estate Prices with Azure Machine Learning
Predicting Moscow Real Estate Prices with Azure Machine Learning
Leo Salemann
 
Predicting Moscow Real Estate Prices with Azure Machine Learning
Predicting Moscow Real Estate Prices with Azure Machine LearningPredicting Moscow Real Estate Prices with Azure Machine Learning
Predicting Moscow Real Estate Prices with Azure Machine Learning
Karunakar Kotha
 
Predicting Moscow Real Estate Prices with Azure Machine Learning
Predicting Moscow Real Estate Prices with Azure Machine LearningPredicting Moscow Real Estate Prices with Azure Machine Learning
Predicting Moscow Real Estate Prices with Azure Machine Learning
Wenfan Xu
 
CVPR2008 tutorial generalized pca
CVPR2008 tutorial generalized pcaCVPR2008 tutorial generalized pca
CVPR2008 tutorial generalized pca
zukun
 
Machine Learning
Machine LearningMachine Learning
Machine Learning
Girish Khanzode
 

Similar to Carolina AI Meetup Nov 2018 (20)

Machine Duping 101: Pwning Deep Learning Systems
Machine Duping 101: Pwning Deep Learning SystemsMachine Duping 101: Pwning Deep Learning Systems
Machine Duping 101: Pwning Deep Learning Systems
 
Machine Learning
Machine LearningMachine Learning
Machine Learning
 
Deep learning - a primer
Deep learning - a primerDeep learning - a primer
Deep learning - a primer
 
Deep learning - a primer
Deep learning - a primerDeep learning - a primer
Deep learning - a primer
 
AI and Deep Learning
AI and Deep Learning AI and Deep Learning
AI and Deep Learning
 
Machine Learning : why we should know and how it works
Machine Learning : why we should know and how it worksMachine Learning : why we should know and how it works
Machine Learning : why we should know and how it works
 
Machine Learning workshop by GDSC Amity University Chhattisgarh
Machine Learning workshop by GDSC Amity University ChhattisgarhMachine Learning workshop by GDSC Amity University Chhattisgarh
Machine Learning workshop by GDSC Amity University Chhattisgarh
 
What is Machine Learning
What is Machine LearningWhat is Machine Learning
What is Machine Learning
 
Intro to Machine Learning by Microsoft Ventures
Intro to Machine Learning by Microsoft VenturesIntro to Machine Learning by Microsoft Ventures
Intro to Machine Learning by Microsoft Ventures
 
Echelon Asia Summit 2017 Startup Academy Workshop
Echelon Asia Summit 2017 Startup Academy WorkshopEchelon Asia Summit 2017 Startup Academy Workshop
Echelon Asia Summit 2017 Startup Academy Workshop
 
Machine learning with R
Machine learning with RMachine learning with R
Machine learning with R
 
Machine learning and linear regression programming
Machine learning and linear regression programmingMachine learning and linear regression programming
Machine learning and linear regression programming
 
DeepLearning
DeepLearningDeepLearning
DeepLearning
 
Getting Started with Keras and TensorFlow - StampedeCon AI Summit 2017
Getting Started with Keras and TensorFlow - StampedeCon AI Summit 2017Getting Started with Keras and TensorFlow - StampedeCon AI Summit 2017
Getting Started with Keras and TensorFlow - StampedeCon AI Summit 2017
 
The Power of Auto ML and How Does it Work
The Power of Auto ML and How Does it WorkThe Power of Auto ML and How Does it Work
The Power of Auto ML and How Does it Work
 
Predicting Moscow Real Estate Prices with Azure Machine Learning
Predicting Moscow Real Estate Prices with Azure Machine LearningPredicting Moscow Real Estate Prices with Azure Machine Learning
Predicting Moscow Real Estate Prices with Azure Machine Learning
 
Predicting Moscow Real Estate Prices with Azure Machine Learning
Predicting Moscow Real Estate Prices with Azure Machine LearningPredicting Moscow Real Estate Prices with Azure Machine Learning
Predicting Moscow Real Estate Prices with Azure Machine Learning
 
Predicting Moscow Real Estate Prices with Azure Machine Learning
Predicting Moscow Real Estate Prices with Azure Machine LearningPredicting Moscow Real Estate Prices with Azure Machine Learning
Predicting Moscow Real Estate Prices with Azure Machine Learning
 
CVPR2008 tutorial generalized pca
CVPR2008 tutorial generalized pcaCVPR2008 tutorial generalized pca
CVPR2008 tutorial generalized pca
 
Machine Learning
Machine LearningMachine Learning
Machine Learning
 

Recently uploaded

storyboard: Victor and Verlin discussing about top hat
storyboard: Victor and Verlin discussing about top hatstoryboard: Victor and Verlin discussing about top hat
storyboard: Victor and Verlin discussing about top hat
LyneSun
 
My storyboard for the short film "Maatla".
My storyboard for the short film "Maatla".My storyboard for the short film "Maatla".
My storyboard for the short film "Maatla".
AlejandroGuarnGutirr
 
Domino Express Storyboard - TV Adv Toys 30"
Domino Express Storyboard - TV Adv Toys 30"Domino Express Storyboard - TV Adv Toys 30"
Domino Express Storyboard - TV Adv Toys 30"
Alessandro Occhipinti
 
My storyboard for a sword fight scene with lightsabers
My storyboard for a sword fight scene with lightsabersMy storyboard for a sword fight scene with lightsabers
My storyboard for a sword fight scene with lightsabers
AlejandroGuarnGutirr
 
Ealing London Independent Photography meeting - June 2024
Ealing London Independent Photography meeting - June 2024Ealing London Independent Photography meeting - June 2024
Ealing London Independent Photography meeting - June 2024
Sean McDonnell
 
Cherries 32 collection of colorful paintings
Cherries 32 collection of colorful paintingsCherries 32 collection of colorful paintings
Cherries 32 collection of colorful paintings
sandamichaela *
 
Complete Lab 123456789123456789123456789
Complete Lab 123456789123456789123456789Complete Lab 123456789123456789123456789
Complete Lab 123456789123456789123456789
vickyvikas51556
 
FinalLessonPlanResponding.docxnknknknknknk
FinalLessonPlanResponding.docxnknknknknknkFinalLessonPlanResponding.docxnknknknknknk
FinalLessonPlanResponding.docxnknknknknknk
abbieharman
 
2024 MATFORCE Youth Poster Contest Winners
2024 MATFORCE Youth Poster Contest Winners2024 MATFORCE Youth Poster Contest Winners
2024 MATFORCE Youth Poster Contest Winners
matforce
 
Codes n Conventions Website Media studies.pptx
Codes n Conventions Website Media studies.pptxCodes n Conventions Website Media studies.pptx
Codes n Conventions Website Media studies.pptx
ZackSpencer3
 
A Brief Introduction About Hadj Ounis
A Brief  Introduction  About  Hadj OunisA Brief  Introduction  About  Hadj Ounis
A Brief Introduction About Hadj Ounis
Hadj Ounis
 
HOW TO USE PINTEREST_by: Clarissa Credito
HOW TO USE PINTEREST_by: Clarissa CreditoHOW TO USE PINTEREST_by: Clarissa Credito
HOW TO USE PINTEREST_by: Clarissa Credito
ClarissaAlanoCredito
 
Colour Theory for Painting - Fine Artist.pdf
Colour Theory for Painting - Fine Artist.pdfColour Theory for Painting - Fine Artist.pdf
Colour Theory for Painting - Fine Artist.pdf
Ketan Naik
 
哪里购买美国乔治城大学毕业证硕士学位证书原版一模一样
哪里购买美国乔治城大学毕业证硕士学位证书原版一模一样哪里购买美国乔治城大学毕业证硕士学位证书原版一模一样
哪里购买美国乔治城大学毕业证硕士学位证书原版一模一样
tc73868
 
Heart Touching Romantic Love Shayari In English with Images
Heart Touching Romantic Love Shayari In English with ImagesHeart Touching Romantic Love Shayari In English with Images
Heart Touching Romantic Love Shayari In English with Images
Short Good Quotes
 
Tibbetts_HappyAwesome_NewArc Sketch to AI
Tibbetts_HappyAwesome_NewArc Sketch to AITibbetts_HappyAwesome_NewArc Sketch to AI
Tibbetts_HappyAwesome_NewArc Sketch to AI
Todd Tibbetts
 
➒➌➎➏➑➐➋➑➐➐ Dpboss Satta Matka Matka Guessing Kalyan Chart Indian Matka Satta ...
➒➌➎➏➑➐➋➑➐➐ Dpboss Satta Matka Matka Guessing Kalyan Chart Indian Matka Satta ...➒➌➎➏➑➐➋➑➐➐ Dpboss Satta Matka Matka Guessing Kalyan Chart Indian Matka Satta ...
➒➌➎➏➑➐➋➑➐➐ Dpboss Satta Matka Matka Guessing Kalyan Chart Indian Matka Satta ...
➒➌➎➏➑➐➋➑➐➐Dpboss Matka Guessing Satta Matka Kalyan Chart Indian Matka
 
In Focus_ The Evolution of Boudoir Photography in NYC.pdf
In Focus_ The Evolution of Boudoir Photography in NYC.pdfIn Focus_ The Evolution of Boudoir Photography in NYC.pdf
In Focus_ The Evolution of Boudoir Photography in NYC.pdf
Boudoir Photography by Your Hollywood Portrait
 
Dino Ranch Storyboard / Kids TV Advertising
Dino Ranch Storyboard / Kids TV AdvertisingDino Ranch Storyboard / Kids TV Advertising
Dino Ranch Storyboard / Kids TV Advertising
Alessandro Occhipinti
 
一比一原版(UniSA毕业证)南澳大学毕业证成绩单如何办理
一比一原版(UniSA毕业证)南澳大学毕业证成绩单如何办理一比一原版(UniSA毕业证)南澳大学毕业证成绩单如何办理
一比一原版(UniSA毕业证)南澳大学毕业证成绩单如何办理
zeyhe
 

Recently uploaded (20)

storyboard: Victor and Verlin discussing about top hat
storyboard: Victor and Verlin discussing about top hatstoryboard: Victor and Verlin discussing about top hat
storyboard: Victor and Verlin discussing about top hat
 
My storyboard for the short film "Maatla".
My storyboard for the short film "Maatla".My storyboard for the short film "Maatla".
My storyboard for the short film "Maatla".
 
Domino Express Storyboard - TV Adv Toys 30"
Domino Express Storyboard - TV Adv Toys 30"Domino Express Storyboard - TV Adv Toys 30"
Domino Express Storyboard - TV Adv Toys 30"
 
My storyboard for a sword fight scene with lightsabers
My storyboard for a sword fight scene with lightsabersMy storyboard for a sword fight scene with lightsabers
My storyboard for a sword fight scene with lightsabers
 
Ealing London Independent Photography meeting - June 2024
Ealing London Independent Photography meeting - June 2024Ealing London Independent Photography meeting - June 2024
Ealing London Independent Photography meeting - June 2024
 
Cherries 32 collection of colorful paintings
Cherries 32 collection of colorful paintingsCherries 32 collection of colorful paintings
Cherries 32 collection of colorful paintings
 
Complete Lab 123456789123456789123456789
Complete Lab 123456789123456789123456789Complete Lab 123456789123456789123456789
Complete Lab 123456789123456789123456789
 
FinalLessonPlanResponding.docxnknknknknknk
FinalLessonPlanResponding.docxnknknknknknkFinalLessonPlanResponding.docxnknknknknknk
FinalLessonPlanResponding.docxnknknknknknk
 
2024 MATFORCE Youth Poster Contest Winners
2024 MATFORCE Youth Poster Contest Winners2024 MATFORCE Youth Poster Contest Winners
2024 MATFORCE Youth Poster Contest Winners
 
Codes n Conventions Website Media studies.pptx
Codes n Conventions Website Media studies.pptxCodes n Conventions Website Media studies.pptx
Codes n Conventions Website Media studies.pptx
 
A Brief Introduction About Hadj Ounis
A Brief  Introduction  About  Hadj OunisA Brief  Introduction  About  Hadj Ounis
A Brief Introduction About Hadj Ounis
 
HOW TO USE PINTEREST_by: Clarissa Credito
HOW TO USE PINTEREST_by: Clarissa CreditoHOW TO USE PINTEREST_by: Clarissa Credito
HOW TO USE PINTEREST_by: Clarissa Credito
 
Colour Theory for Painting - Fine Artist.pdf
Colour Theory for Painting - Fine Artist.pdfColour Theory for Painting - Fine Artist.pdf
Colour Theory for Painting - Fine Artist.pdf
 
哪里购买美国乔治城大学毕业证硕士学位证书原版一模一样
哪里购买美国乔治城大学毕业证硕士学位证书原版一模一样哪里购买美国乔治城大学毕业证硕士学位证书原版一模一样
哪里购买美国乔治城大学毕业证硕士学位证书原版一模一样
 
Heart Touching Romantic Love Shayari In English with Images
Heart Touching Romantic Love Shayari In English with ImagesHeart Touching Romantic Love Shayari In English with Images
Heart Touching Romantic Love Shayari In English with Images
 
Tibbetts_HappyAwesome_NewArc Sketch to AI
Tibbetts_HappyAwesome_NewArc Sketch to AITibbetts_HappyAwesome_NewArc Sketch to AI
Tibbetts_HappyAwesome_NewArc Sketch to AI
 
➒➌➎➏➑➐➋➑➐➐ Dpboss Satta Matka Matka Guessing Kalyan Chart Indian Matka Satta ...
➒➌➎➏➑➐➋➑➐➐ Dpboss Satta Matka Matka Guessing Kalyan Chart Indian Matka Satta ...➒➌➎➏➑➐➋➑➐➐ Dpboss Satta Matka Matka Guessing Kalyan Chart Indian Matka Satta ...
➒➌➎➏➑➐➋➑➐➐ Dpboss Satta Matka Matka Guessing Kalyan Chart Indian Matka Satta ...
 
In Focus_ The Evolution of Boudoir Photography in NYC.pdf
In Focus_ The Evolution of Boudoir Photography in NYC.pdfIn Focus_ The Evolution of Boudoir Photography in NYC.pdf
In Focus_ The Evolution of Boudoir Photography in NYC.pdf
 
Dino Ranch Storyboard / Kids TV Advertising
Dino Ranch Storyboard / Kids TV AdvertisingDino Ranch Storyboard / Kids TV Advertising
Dino Ranch Storyboard / Kids TV Advertising
 
一比一原版(UniSA毕业证)南澳大学毕业证成绩单如何办理
一比一原版(UniSA毕业证)南澳大学毕业证成绩单如何办理一比一原版(UniSA毕业证)南澳大学毕业证成绩单如何办理
一比一原版(UniSA毕业证)南澳大学毕业证成绩单如何办理
 

Carolina AI Meetup Nov 2018

  • 2. OUTLINE • What is Machine Learning? • Applications in Machine Learning • (The Machine Learning) Model • Machine Learning Models in Action • Training Data • Model / Data Considerations • Models • DecisionTree • Random Forest • Clustering • Linear Models • SupportVector Machines (SVM) • Artificial Neural Networks • Deep Learning (CNN) • Reinforcement Learning
  • 3. WHAT IS MACHINE LEARNING? “Field of study that gives computers the ability to learn without being explicitly programmed.” - Arthur Samuel A computer program is said to learn from experience E with respect to some taskT and some performance measure P, if its performance onT, as measured by P, improves with experience E. -Tom Mitchell1959 1998
  • 4. WHAT IS MACHINE LEARNING?
  • 5. APPLICATIONS INTHE MODERN WORLDAPPLICATIONS INTHE MODERN WORLD Optical Character Recognition Recommendation Engines Facial Recognition Autonomous Vehicles Personal Assistants / Chat Bots
  • 6.
  • 7. MODEL A REPRESENTATION OF A REALWORLD PROCESS Water Cycle
  • 8. MODEL A REPRESENTATION OF A REALWORLD PROCESS Water Cycle Evolution
  • 9. MODEL A REPRESENTATION OF A REALWORLD PROCESS Water Cycle Evolution Neuron-McCulloch & Pitts Model, 1943
  • 10. MACHINE LEARNING MODELS IN ACTION Untrained Model(Old) DATA Trained Model New Data Info? Prediction? Decision? Expert Knowledge
  • 11. TRAINING DATA Feature =Variable = Predictor Objective Measurement Height (in) Weight (lb) Color Claws retract Class 11.2 10.1 black yes cat 23.1 45.2 black/white no dog 13.0 20.1 black/white yes cat 9.7 7.2 white yes cat … … … … …
  • 12. TRAINING DATA Feature =Variable = Predictor Objective Measurement Height (in) Weight (lb) Color Claws retract Class 11.2 10.1 black yes cat 23.1 45.2 black/white no dog 13.0 20.1 black/white yes cat 9.7 7.2 white yes cat … … … … …
  • 13. TRAINING DATA Feature =Variable = Predictor Objective Measurement Height (in) Weight (lb) Color Claws retract Class 11.2 10.1 black yes 23.1 45.2 black/white no dog 13.0 20.1 black/white yes 9.7 7.2 white yes cat … … … … …
  • 14. TESTING DATA (NO PEEKING!) Training and testing sets must ALWAYS be disjoint • Cross-validation • Leave-one-out • OOB (Out-of-bag for ensembles)
  • 15. MODEL/DATA CONSIDERATIONS (RELEVANT TO MODEL SELECTION) Each model can/cannot handle certain data characteristics / analysis needs • Supervised vs. Unsupervised data? • Class Imbalance (200 cats vs. 3 dogs) • 2-class vs. Multiclass (say 200 cats, 146 dogs, 25 sugar gliders, 5 platypuses) • Scale issues (see Distance-based Clustering; Normalization / Standardization) • FeatureType (Categorical, Continuous, etc) • Dimensionality (# of features / measurements) • Cost Sensitivity (Miss / False Alarm – can the model adjust?) • Propensity to Overtrain (fitting to noise – see Bias vs.Variance)? • Need to estimate uncertainty? • Ability to adapt to changing conditions (parameters)? • Robustness to sparse data (parameter estimation)?
  • 16. DECISIONTREE 1) At each node, a question is asked about a specific feature 2) The answer directs data left/right 3) Decision trees must be pruned to prevent overtraining
  • 17. RANDOM FOREST Random Forest is an ENSEMBLE of DecisionTrees RANDOM FOREST
  • 18. RANDOM FOREST Random Forest is an ENSEMBLE of DecisionTrees Node Splits (Training) • Bagging (resampled data for each tree) • “Best” univariate split on random subspace (subset of all features) • Gini Impurity • Leaf nodes are class homogeneousLeo Breiman
  • 19. RANDOM FOREST Random Forest is an ENSEMBLE of DecisionTrees Leo Breiman Classification 1) Samples propagate through each tree 2) Tree “votes” for a class based on leaf node 3) Final decision based on class conditional probability
  • 21. CLUSTERING KeyVariants K-means: point-to-cluster mean distance Mean-Shift: hill-climbing to max density
  • 22. CLUSTERING KeyVariants K-means: point-to-cluster mean distance Mean-Shift: hill-climbing to max density DBSCAN: epsilon neighborhood
  • 23. CLUSTERING KeyVariants K-means: point-to-cluster mean distance Mean-Shift: hill-climbing to max density DBSCAN: epsilon neighborhood Gaussian Mixture Models: Gaussian assumption
  • 24. CLUSTERING KeyVariants K-means: point-to-cluster mean distance Mean-Shift: hill-climbing to max density Hierarchical Clustering DBSCAN: epsilon neighborhood Gaussian Mixture Models: Gaussian assumption
  • 25. LINEAR MODELS Linear Discriminant Analysis Simple Linear Regression (Ronald) Fisher’s LDA
  • 26. SUPPORTVECTOR MACHINES (SVM) Maps linearly nonseparable data to a higher dimension Kernel trick makes this mapping more efficient Also: sub-gradient descent, coordinate descent
  • 27. SUPPORTVECTOR MACHINES (SVM) Support vectors in the feature space used for classification Support vectors are determined by the most difficult points to classify…
  • 28. ARTIFICIAL NEURAL NETWORKS Recall the original model of the neuron…
  • 29. ARTIFICIAL NEURAL NETWORKS Input Layer Hidden Layers Output Layer Feedforward (forward processing) • Each arrow represents a weight • Hidden & output nodes “process” input values/weights Backpropagation (of errors) • Allow specification of desired output • Minimize loss function w11 w12 w41 f31 f41 f51
  • 30. ARTIFICIAL NEURAL NETWORKS Input Layer Hidden Layers Output Layer w11 w12 w41 f31 f41 f51
  • 31. DEEP LEARNING (CONVOLUTIONAL NN) From the Latin convolvere,“to convolve” means to roll together We convolve an image with multiple kernels (filters) at each layer
  • 32. DEEP LEARNING (CONVOLUTIONAL NN) Each layer of the network learns different features of the image
  • 34. REINFORCEMENT LEARNING A reward-driven approach for a machine to “self-learn” • At each step, the agent takes an action based on environment state • The agent receives a reward based upon the new state (post-action) • The agent’s goal is to maximize his reward
  • 35. REINFORCEMENT LEARNING Donald Michie creates MENACE, 1963 (Machine Educable Noughts And Crosses Engine) MENACE learned to play TicTacToe using stacks of matchboxes
  • 36. REINFORCEMENT LEARNING Q(uality)-Learning – values-based; environment may be unknown