machine learning lecture introduction · what is machine learning? 1) machine learning is an...
TRANSCRIPT
![Page 1: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/1.jpg)
Machine Learning
Introduction
Marc ToussaintU Stuttgart
![Page 2: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/2.jpg)
What is Machine Learning?1) Machine Learning is an approach to understand learning
by building learning systems(“synthetic” approach to a science of learning)
2) A long list of methods/algorithms for different data anlysis problems— in sciences— in commerce
3) A framework to develop your own learning algorithms/methods
4) Machine Learning = information theory/statistics + computer science
2/25
![Page 3: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/3.jpg)
What is Machine Learning?1) Machine Learning is an approach to understand learning
by building learning systems(“synthetic” approach to a science of learning)
2) A long list of methods/algorithms for different data anlysis problems— in sciences— in commerce
3) A framework to develop your own learning algorithms/methods
4) Machine Learning = information theory/statistics + computer science
2/25
![Page 4: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/4.jpg)
What is Machine Learning?1) Machine Learning is an approach to understand learning
by building learning systems(“synthetic” approach to a science of learning)
2) A long list of methods/algorithms for different data anlysis problems— in sciences— in commerce
3) A framework to develop your own learning algorithms/methods
4) Machine Learning = information theory/statistics + computer science
2/25
![Page 5: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/5.jpg)
What is Machine Learning?1) Machine Learning is an approach to understand learning
by building learning systems(“synthetic” approach to a science of learning)
2) A long list of methods/algorithms for different data anlysis problems— in sciences— in commerce
3) A framework to develop your own learning algorithms/methods
4) Machine Learning = information theory/statistics + computer science
2/25
![Page 6: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/6.jpg)
Examples for ML applications...
3/25
![Page 7: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/7.jpg)
Face recognition
keypoints
eigenfaces
(e.g., Viola & Jones)
4/25
![Page 8: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/8.jpg)
Hand-written digit recognition (US postal data)
(e.g., Yann LeCun)
5/25
![Page 9: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/9.jpg)
Gene annotation
(Gunnar Ratsch, Tubingen, mGene Project)
6/25
![Page 10: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/10.jpg)
Speech recognition
(This is the basis of all commercial products)
7/25
![Page 11: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/11.jpg)
Spam filters
8/25
![Page 12: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/12.jpg)
• More examples:
– Google’s (and many others’) analysis of user preferences
– Medical diagnosis
• Machine Learning became an important technologyin science as well as commerce
9/25
![Page 13: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/13.jpg)
Examples of ML for behavior...
10/25
![Page 14: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/14.jpg)
Learing motor skills
(around 2000, by Schaal, Atkeson, Vijayakumar)
11/25
![Page 15: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/15.jpg)
Learning to walk
(Rus Tedrake et al.)
12/25
![Page 16: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/16.jpg)
Learning effects of actions
(Tobias Lang & M Toussaint)
13/25
![Page 17: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/17.jpg)
Types of ML
• Supervised learning: learn from “labelled” data {(xi, yi)}Ni=1
Unsupervised learning: learn from “unlabelled” data {xi}Ni=0 onlySemi-supervised learning: many unlabelled data, few labelled data
• Reinforcement learning: learn from data {(st, at, rt, st+1)}– learn a predictive model (s, a) 7→ s′
– learn to predict reward (s, a) 7→ r
– learn a behavior s 7→ a that maximizes reward
14/25
![Page 18: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/18.jpg)
Types of ML
• Supervised learning: learn from “labelled” data {(xi, yi)}Ni=1
Unsupervised learning: learn from “unlabelled” data {xi}Ni=0 onlySemi-supervised learning: many unlabelled data, few labelled data
• Reinforcement learning: learn from data {(st, at, rt, st+1)}– learn a predictive model (s, a) 7→ s′
– learn to predict reward (s, a) 7→ r
– learn a behavior s 7→ a that maximizes reward
14/25
![Page 19: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/19.jpg)
Types of ML
• Supervised learning: learn from “labelled” data {(xi, yi)}Ni=1
Unsupervised learning: learn from “unlabelled” data {xi}Ni=0 onlySemi-supervised learning: many unlabelled data, few labelled data
• Reinforcement learning: learn from data {(st, at, rt, st+1)}– learn a predictive model (s, a) 7→ s′
– learn to predict reward (s, a) 7→ r
– learn a behavior s 7→ a that maximizes reward
14/25
![Page 20: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/20.jpg)
Organization of this lecture• Part 1: The Basis
• Basic regression & classification
• Part 2: The Breadth of ML ideas• PCA, PLS• Local & lazy learning• Combining weak learners: boosting, decision trees & stumps• Other loss functions & Sparse approximations: SVMs• Deep learning
• Part 3: In Depth Topics• Bayesian Learning, Bayesian Ridge/Logistic Regression
Gaussian Processes GP classification• Active Learning
Recommender Systems
• Missing:– Neural Networks– Graphical Models & structure learning– un-, semi-supervised learning 15/25
![Page 21: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/21.jpg)
• Is this a theoretical or practical course?
Neither alone.
• The goal is to teach how to design good learning algorithms
data↓
modelling [requires theory & practise]↓
algorithms [requires practise & theory]↓
testing, problem identification, restart
16/25
![Page 22: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/22.jpg)
• Is this a theoretical or practical course?
Neither alone.
• The goal is to teach how to design good learning algorithms
data↓
modelling [requires theory & practise]↓
algorithms [requires practise & theory]↓
testing, problem identification, restart
16/25
![Page 23: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/23.jpg)
Basic regression & classification
-1.5
-1
-0.5
0
0.5
1
1.5
-3 -2 -1 0 1 2 3
(MT/plot.h -> gnuplot pipe)
'train' us 1:2'model' us 1:2
-2
-1
0
1
2
3
-2 -1 0 1 2 3
(MT/plot.h -> gnuplot pipe)
traindecision boundary
• Regression: map input x to continuous value y ∈ RClassification: map input x to one of M classes y ∈ {1, 2, ..,M}
17/25
![Page 24: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/24.jpg)
Basic regression & classification
• A must-know!
• High practical relevance for applications
• Focus on linear methods on non-linear features, regularization,cross-validation
“linear|polynomial|Kernel Ridge|Lasso Regression|Classification”
• Relations to SVM, GPs, feature selection
18/25
![Page 25: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/25.jpg)
Bayesian Modelling
• Mr. Holmes lives in Los Angeles. One morning when Holmes leaveshis house, he realizes that his grass is wet. Is it due to rain, or has heforgotten to turn off his sprinkler?
• Holmes checks Watsons grass, and finds it is also wet. What does thatimply on rain vs. sprinkler?
Watson Holmes
P(W=yes|R=yes)=1.0
P(W=yes|R=no)=0.2 P(H=yes|R=yes,S=yes)=1.0
P(H=yes|R=yes,S=no)=1.0
P(H=yes|R=no,S=yes)=0.9
P(H=yes|R=no,S=no)=0.0
Rain
P(R=yes)=0.2
Sprinkler
P(S=yes)=0.1
⇐⇒ P (H,W,S,R) = P (H|S,R) P (W |R) P (S) P (R)19/25
![Page 26: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/26.jpg)
Bayesian modelling
• Fundamental view on information processing and learning
• Provides general tools for formulating structured probabilistic models(e.g., latent variables, mixtures, hierarchical, deep models)→ a framework for formulating novel learning algorithms
• Bayesian view on linear models + regularization– regularization↔ prior, “error”↔ likelihood
20/25
![Page 27: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/27.jpg)
Reinforcement Learning (another course..)[PacMan]
I Szita, A Lorincz: Learning to Play Using Low-Complexity Rule-BasedPolicies: Illustrations through Ms. Pac-Man. JAIR 2007.
21/25
![Page 28: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/28.jpg)
Reinforcement Learning (another course..)
• Behavior!, learning to act
• Basic RL methods (Temporal Difference, Q-learning, traces)
• Regression in RL, Bayesian methods in RL
• Applications
22/25
![Page 29: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/29.jpg)
Books
Springer Series in Statistics
Trevor HastieRobert TibshiraniJerome Friedman
Springer Series in Statistics
The Elements ofStatistical LearningData Mining, Inference, and Prediction
The Elements of Statistical Learning
During the past decade there has been an explosion in computation and information tech-nology. With it have come vast amounts of data in a variety of fields such as medicine, biolo-gy, finance, and marketing. The challenge of understanding these data has led to the devel-opment of new tools in the field of statistics, and spawned new areas such as data mining,machine learning, and bioinformatics. Many of these tools have common underpinnings butare often expressed with different terminology. This book describes the important ideas inthese areas in a common conceptual framework. While the approach is statistical, theemphasis is on concepts rather than mathematics. Many examples are given, with a liberaluse of color graphics. It should be a valuable resource for statisticians and anyone interestedin data mining in science or industry. The book’s coverage is broad, from supervised learning(prediction) to unsupervised learning. The many topics include neural networks, supportvector machines, classification trees and boosting—the first comprehensive treatment of thistopic in any book.
This major new edition features many topics not covered in the original, including graphicalmodels, random forests, ensemble methods, least angle regression & path algorithms for thelasso, non-negative matrix factorization, and spectral clustering. There is also a chapter onmethods for “wide” data (p bigger than n), including multiple testing and false discovery rates.
Trevor Hastie, Robert Tibshirani, and Jerome Friedman are professors of statistics atStanford University. They are prominent researchers in this area: Hastie and Tibshiranideveloped generalized additive models and wrote a popular book of that title. Hastie co-developed much of the statistical modeling software and environment in R/S-PLUS andinvented principal curves and surfaces. Tibshirani proposed the lasso and is co-author of thevery successful An Introduction to the Bootstrap. Friedman is the co-inventor of many data-mining tools including CART, MARS, projection pursuit and gradient boosting.
› springer.com
S T A T I S T I C S
----
Trevor Hastie • Robert Tibshirani • Jerome FriedmanThe Elements of Statictical Learning
Hastie • Tibshirani • Friedman
Second Edition
Trevor Hastie, Robert Tibshirani andJerome Friedman: The Elements ofStatistical Learning: Data Mining, In-ference, and Prediction Springer, Sec-ond Edition, 2009.http://www-stat.stanford.edu/
~tibs/ElemStatLearn/
(recommended: read introductorychapter)
(this course will not go to the full depth in math of Hastie et al.)
23/25
![Page 30: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/30.jpg)
Books
Bishop, C. M.: Pattern Recognitionand Machine Learning.Springer, 2006http://research.microsoft.com/
en-us/um/people/cmbishop/prml/
(some chapters are fully online)
24/25
![Page 31: Machine Learning Lecture Introduction · What is Machine Learning? 1) Machine Learning is an approach to understand learning by building learning systems (“synthetic” approach](https://reader030.vdocuments.us/reader030/viewer/2022021702/5c98114209d3f2d8238d2443/html5/thumbnails/31.jpg)
Organisation
• Vorlesungs-Webpage:http://ipvs.informatik.uni-stuttgart.de/mlr/marc/teaching/13-MachineLearning/
– Slides, Ubungen & Software (C++)– Links zu Buchern und anderen Ressourcen
• Sekretariat/Organisatorische Fragen:Carola Stahl, [email protected], Raum 2.217
• 2 geplante Ubungen: Dienstag 14:00-15:30 & 15:45-17:15, 0.453
• Regelung zu Ubungen:– Bearbeitung der Ubungen ist wichtig!– Zu Beginn jeder Ubung in Liste eintragen:
– Teilnahme– Welche Aufgaben wurden bearbeitet
– Zufallige Auswahl zur Prasentation der Losung– 50% bearbeitete Aufgaben notwendig fur aktive Teilnahme
25/25