Semi-Supervised Learning

Lukas Tencer
Lukas TencerSoftware Engineer um Synchromedia
Semi-Supervised Learning 
Lukas Tencer 
PhD student @ ETS
Motivation
Image Similarity 
- Domain of origin 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Face Recognition 
- Cross-race effect 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Motivation in Machine Learning 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Motivation in Machine Learning 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Methodology
When to use Semi-Supervised Learning? 
• Labelled data is hard to get and expensive 
– Speech analysis: 
• Switchboard dataset 
• 400 hours annotation time for 1 hour of speech 
– Natural Language Processing 
• Penn Chinese Treebank 
• 2 Years for 4000 sentences 
– Medical Application 
• Require experts opinion which might not be unique 
• Unlabelled data is cheap 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Types of Semi-Supervised Leaning 
• Transductive Learning 
– Does not generalize to unseen data 
– Produces labels only for the data at training time 
• 1. Assume labels 
• 2. Train classifier on assumed labels 
• Inductive Learning 
– Does generalize to unseen data 
– Not only produces labels, but also the final classifier 
– Manifold Assumption 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Selected Semi-Supervised Algorithms 
• Self-Training 
• Help-Training 
• Transductive SVM (S3VM) 
• Multiview Algorithms 
• Graph-Based Algorithms 
• Generative Models 
• ……. 
….. 
… 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Self-Training 
• The Idea: If I am highly confident in a label of examples, I 
am right 
• Given Training set 푇 = {푥푖 }, and unlabelled set 푈 = {푢푗 } 
1. Train 푓 on 푇 
2. Get predictions 푃 = 푓(푈) 
3. If 푃푖 > 훼 then add (푥, 푓(푥)) to 푇 
4. Retrain 푓 on 푇 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Self-Training 
• Advantages: 
– Very simple and fast method 
– Frequently used in NLP 
• Disadvantages: 
– Amplifies noise in labeled data 
– Requires explicit definition of 푃 푦 푥 
– Hard to implement for discriminative classifiers (SVM) 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Self-Training 
1. Naïve Bayes Classifier on Bag-of-Visual-Word for 2 Classes 
2. Classify Unlabelled Data base on Learned Classifier 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Self-Training 
3. Add the most confident images to the training set 
4. Retrain and repeat 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Help-Training 
• The Challenge: How to make Self-Training work for 
Discriminative Classifiers (SVM) ? 
• The Idea: Train Generative Help Classifier to get 푝(푦|푥) 
• Given Training set 푇 = {푥푖 }, unlabelled set 푈 = {푢푗 }, and 
generative classifier 푔 and discriminative classifier 푓 
1. Train 푓 and 푔 on 푇 
2. Get predictions 푃푔 = 푔(푈) and 푃푓 = 푓(푈) 
3. If 푃푔,푖 > 훼 then add (푥, 푓(푥)) to 푇 
4. Reduce the value of 훼 if |푃푔,푖 > 훼| = 0 
5. Retrain 푓 and 푔 on 푇 until 푈 = 0 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Transductive SVM (S3VM) 
• The Idea: Find largest margin classifier, such that, 
unlabelled data are outside of the margin as much as 
possible, use regularization over unlabelled data 
• Given Training set 푇 = {푥푖 }, and unlabelled set 푈 = {푢푗 } 
1. Find all possible labelings 푈1 ⋯ 푈푛 on 푈 
2. For each 푇 푘 = 푇 ∪ 푈푘 train a standard SVM 
3. Choose SVM with largest margins 
• What is the catch? 
• NP hard problem, fortunately approximations exist 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Transductive SVM (S3VM) 
• Solving non-convex optimization problem: 
퐽 휃 = 
• Methods: 
1 
2 
푤 2 + 푐1 
푥푖∈푇 
퐿(푦푖푓휃 (푥푖 )) + 푐2 
– Local Combinatorial Search 
– Standard unconstrained optimization solvers (CG, BFGS…) 
– Continuation Methods 
– Concave-Convex procedure (CCCP) 
– Branch and Bound 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data :: 
푥푖∈푈 
퐿( 푓휃 (푥푖 ) )
Transductive SVM (S3VM) 
• Advantages: 
– Can be used with any SVM 
– Clear optimization criterion, mathematically well 
formulated 
• Disadvantages: 
– Hard to optimize 
– Prone to local minima – non convex 
– Only small gain given modest assumptions 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Multiview Algorithms 
• The Idea: Train 2 classifiers on 2 disjoint sets of features, 
then let each classifier label unlabelled examples and 
teach the other classifier 
• Given Training set 푇 = {푥푖 }, and unlabelled set 푈 = {푢푗 } 
1. Split 푇 into 푇1 and 푇2 on the feature dimension 
2. Train 푓1 on 푇1 and 푓1 on 푇2 
3. Get predictions 푃1 = 푓1(푈) and 푃2 = 푓2(푈) 
4. Add: top 푘 from 푃1 to 푇2; top 푘 from 푃1 to 푇1 
5. Repeat until 푈 = 0 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Multiview Algorithms 
• Application: Web-page Topic Classification 
– 1. Classifier for Images; 2. Classifier for Text 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Multiview Algorithms 
• Advantages: 
– Simple Method applicable to any classifier 
– Can correct mistakes in classification between the 2 
classifiers 
• Disadvantages: 
– Assumes conditional independence between features 
– Natural split may not exist 
– Artificial split may be complicated if only few eatures 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Graph-Based Algorithms 
• The Idea: Create a connected graph from labelled and 
unlabelled examples, propagate labels over the graph 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Graph-Based Algorithms 
• Advantages: 
– Great performance if graph fits the tasks 
– Can be used in combination with any model 
– Explicit mathematical formulation 
• Disadvantages: 
– Problem if graph does not fit the task 
– Hard to construct graph in sparse spaces 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Generative Models 
• The Idea: Assume distribution using labelled data, update 
using unlabelled data 
• Simple models is: 
GMM + EM 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Generative Models 
• Advantages: 
– Nice probabilistic framework 
– Instead of EM you can go full Bayesian and include 
prior with MAP 
• Disadvantages: 
– EM find only local minima 
– Makes strong assumptions about class distributions 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
What could go wrong? 
• Semi-Supervised Learning make a lot of assumptions 
– Smoothness 
– Clusters 
– Manifolds 
• Some techniques (Co-Training) require very specific 
setup 
• Frequently problem with noisy labels 
• There is no free lunch 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
There is much more out there 
• Structural Learning 
• Co-EM 
• Tri-Training 
• Co-Boosting 
• Unsupervised pretraining – deep learning 
• Transductive Inference 
• Universum Learning 
• Active Learning + Semi-Supervised Learning 
• ……. 
• ….. 
• … 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data :: 
My work
Demo
Conclusion 
• Play with Semi-Supervised Learning 
• Basic methods are vary simple to implement and can give 
you up to 5 to 10% accuracy 
• You can cheat at competitions by using unlabelled data, 
often no assumption is made about external data 
• Be careful when running Semi-Supervised Learning in 
production environment, keep an eye on your algorithm 
• If running in production, be aware that data patterns 
change and old assumptions about labels may screw up 
you new unlabelled data 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
Some more resources 
Videos to watch: 
Semisupervised Learning Approaches – Tom Mitchell CMU : 
http://videolectures.net/mlas06_mitchell_sla/ 
MLSS 2012 Graph based semi-supervised learning - Zoubin 
Ghahramani Cambridge : 
https://www.youtube.com/watch?v=HZQOvm0fkLA 
Books to read: 
• Semi-Supervised Learning – Chapelle, Schölkopf, Zien 
• Introduction to Semi-Supervised Learning - Zhu, Oldberg, 
Brachman, Dietterich 
:: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
THANKS FOR YOUR TIME 
Lukas Tencer 
lukas.tencer@gmail.com 
http://lukastencer.github.io/ 
https://github.com/lukastencer 
https://twitter.com/lukastencer 
Graduating August 2015, looking for ML and DS opportunities
1 von 31

Recomendados

Support Vector Machines von
Support Vector MachinesSupport Vector Machines
Support Vector Machinesnextlib
19.9K views56 Folien
Ensemble learning von
Ensemble learningEnsemble learning
Ensemble learningHaris Jamil
8.6K views14 Folien
Classification and Regression von
Classification and RegressionClassification and Regression
Classification and RegressionMegha Sharma
1.9K views12 Folien
Semi-supervised Learning von
Semi-supervised LearningSemi-supervised Learning
Semi-supervised Learningbutest
4.2K views29 Folien
Support vector machine-SVM's von
Support vector machine-SVM'sSupport vector machine-SVM's
Support vector machine-SVM'sAnudeep Chowdary Kamepalli
487 views14 Folien
Support vector machine von
Support vector machineSupport vector machine
Support vector machineSomnathMore3
609 views20 Folien

Más contenido relacionado

Was ist angesagt?

Semi supervised learning machine learning made simple von
Semi supervised learning  machine learning made simpleSemi supervised learning  machine learning made simple
Semi supervised learning machine learning made simpleDevansh16
180 views12 Folien
Machine learning clustering von
Machine learning clusteringMachine learning clustering
Machine learning clusteringCosmoAIMS Bassett
4.8K views15 Folien
Introduction to ML (Machine Learning) von
Introduction to ML (Machine Learning)Introduction to ML (Machine Learning)
Introduction to ML (Machine Learning)SwatiTripathi44
2K views31 Folien
supervised learning von
supervised learningsupervised learning
supervised learningAmar Tripathi
24.9K views20 Folien
Machine learning seminar ppt von
Machine learning seminar pptMachine learning seminar ppt
Machine learning seminar pptRAHUL DANGWAL
7.2K views17 Folien

Was ist angesagt?(20)

Semi supervised learning machine learning made simple von Devansh16
Semi supervised learning  machine learning made simpleSemi supervised learning  machine learning made simple
Semi supervised learning machine learning made simple
Devansh16180 views
Introduction to ML (Machine Learning) von SwatiTripathi44
Introduction to ML (Machine Learning)Introduction to ML (Machine Learning)
Introduction to ML (Machine Learning)
SwatiTripathi442K views
Machine learning seminar ppt von RAHUL DANGWAL
Machine learning seminar pptMachine learning seminar ppt
Machine learning seminar ppt
RAHUL DANGWAL7.2K views
Supervised and unsupervised learning von AmAn Singh
Supervised and unsupervised learningSupervised and unsupervised learning
Supervised and unsupervised learning
AmAn Singh320 views
Support Vector Machine ppt presentation von AyanaRukasar
Support Vector Machine ppt presentationSupport Vector Machine ppt presentation
Support Vector Machine ppt presentation
AyanaRukasar2K views
Presentation on supervised learning von Tonmoy Bhagawati
Presentation on supervised learningPresentation on supervised learning
Presentation on supervised learning
Tonmoy Bhagawati24.2K views
Understanding Bagging and Boosting von Mohit Rajput
Understanding Bagging and BoostingUnderstanding Bagging and Boosting
Understanding Bagging and Boosting
Mohit Rajput2.7K views
Machine learning ppt. von ASHOK KUMAR
Machine learning ppt.Machine learning ppt.
Machine learning ppt.
ASHOK KUMAR1.5K views
Support Vector Machines for Classification von Prakash Pimpale
Support Vector Machines for ClassificationSupport Vector Machines for Classification
Support Vector Machines for Classification
Prakash Pimpale35.8K views
Reinforcement Learning Q-Learning von Melaku Eneayehu
Reinforcement Learning   Q-Learning Reinforcement Learning   Q-Learning
Reinforcement Learning Q-Learning
Melaku Eneayehu4.9K views
Bias and variance trade off von VARUN KUMAR
Bias and variance trade offBias and variance trade off
Bias and variance trade off
VARUN KUMAR702 views
Introduction to Machine Learning von Lior Rokach
Introduction to Machine LearningIntroduction to Machine Learning
Introduction to Machine Learning
Lior Rokach283.4K views
Supervised and unsupervised learning von Paras Kohli
Supervised and unsupervised learningSupervised and unsupervised learning
Supervised and unsupervised learning
Paras Kohli9.7K views
Feature selection concepts and methods von Reza Ramezani
Feature selection concepts and methodsFeature selection concepts and methods
Feature selection concepts and methods
Reza Ramezani9K views
Machine Learning With Logistic Regression von Knoldus Inc.
Machine Learning  With Logistic RegressionMachine Learning  With Logistic Regression
Machine Learning With Logistic Regression
Knoldus Inc.5.6K views

Similar a Semi-Supervised Learning

How Machine Learning Helps Organizations to Work More Efficiently? von
How Machine Learning Helps Organizations to Work More Efficiently?How Machine Learning Helps Organizations to Work More Efficiently?
How Machine Learning Helps Organizations to Work More Efficiently?Tuan Yang
2K views54 Folien
intership summary von
intership summaryintership summary
intership summaryJunting Ma
200 views44 Folien
1. Intro DS.pptx von
1. Intro DS.pptx1. Intro DS.pptx
1. Intro DS.pptxAnusuya123
8 views15 Folien
Introduction to Deep learning von
Introduction to Deep learningIntroduction to Deep learning
Introduction to Deep learningMassimiliano Ruocco
2.7K views33 Folien
EssentialsOfMachineLearning.pdf von
EssentialsOfMachineLearning.pdfEssentialsOfMachineLearning.pdf
EssentialsOfMachineLearning.pdfAnkita Tiwari
15 views65 Folien
Hacking Predictive Modeling - RoadSec 2018 von
Hacking Predictive Modeling - RoadSec 2018Hacking Predictive Modeling - RoadSec 2018
Hacking Predictive Modeling - RoadSec 2018HJ van Veen
1.2K views74 Folien

Similar a Semi-Supervised Learning(20)

How Machine Learning Helps Organizations to Work More Efficiently? von Tuan Yang
How Machine Learning Helps Organizations to Work More Efficiently?How Machine Learning Helps Organizations to Work More Efficiently?
How Machine Learning Helps Organizations to Work More Efficiently?
Tuan Yang2K views
intership summary von Junting Ma
intership summaryintership summary
intership summary
Junting Ma200 views
EssentialsOfMachineLearning.pdf von Ankita Tiwari
EssentialsOfMachineLearning.pdfEssentialsOfMachineLearning.pdf
EssentialsOfMachineLearning.pdf
Ankita Tiwari15 views
Hacking Predictive Modeling - RoadSec 2018 von HJ van Veen
Hacking Predictive Modeling - RoadSec 2018Hacking Predictive Modeling - RoadSec 2018
Hacking Predictive Modeling - RoadSec 2018
HJ van Veen1.2K views
in5490-classification (1).pptx von MonicaTimber
in5490-classification (1).pptxin5490-classification (1).pptx
in5490-classification (1).pptx
MonicaTimber34 views
Supervised machine learning algorithms(strengths and weaknesses) von MonarchSaha
Supervised machine learning algorithms(strengths and weaknesses)Supervised machine learning algorithms(strengths and weaknesses)
Supervised machine learning algorithms(strengths and weaknesses)
MonarchSaha251 views
Enriching Solr with Deep Learning for a Question Answering System - Sanket Sh... von Lucidworks
Enriching Solr with Deep Learning for a Question Answering System - Sanket Sh...Enriching Solr with Deep Learning for a Question Answering System - Sanket Sh...
Enriching Solr with Deep Learning for a Question Answering System - Sanket Sh...
Lucidworks564 views
04-Data-Analysis-Overview.pptx von Shree Shree
04-Data-Analysis-Overview.pptx04-Data-Analysis-Overview.pptx
04-Data-Analysis-Overview.pptx
Shree Shree16 views
Machine Learning, Deep Learning and Data Analysis Introduction von Te-Yen Liu
Machine Learning, Deep Learning and Data Analysis IntroductionMachine Learning, Deep Learning and Data Analysis Introduction
Machine Learning, Deep Learning and Data Analysis Introduction
Te-Yen Liu6.3K views
Hyperparameter Tuning von Jon Lederman
Hyperparameter TuningHyperparameter Tuning
Hyperparameter Tuning
Jon Lederman2.9K views
An Introduction to Deep Learning von milad abbasi
An Introduction to Deep LearningAn Introduction to Deep Learning
An Introduction to Deep Learning
milad abbasi63 views
Introduction to Deep Learning von Mehrnaz Faraz
Introduction to Deep LearningIntroduction to Deep Learning
Introduction to Deep Learning
Mehrnaz Faraz198 views
LETS PUBLISH WITH MORE RELIABLE & PRESENTABLE MODELLING.pptx von shamsul2010
LETS PUBLISH WITH MORE RELIABLE & PRESENTABLE MODELLING.pptxLETS PUBLISH WITH MORE RELIABLE & PRESENTABLE MODELLING.pptx
LETS PUBLISH WITH MORE RELIABLE & PRESENTABLE MODELLING.pptx
shamsul20103 views

Más de Lukas Tencer

ICRA: Intelligent Platform for Collaboration and Interaction von
ICRA: Intelligent Platform for Collaboration and InteractionICRA: Intelligent Platform for Collaboration and Interaction
ICRA: Intelligent Platform for Collaboration and InteractionLukas Tencer
630 views68 Folien
Introduction to Probability von
Introduction to ProbabilityIntroduction to Probability
Introduction to ProbabilityLukas Tencer
651 views29 Folien
Common Probability Distibution von
Common Probability DistibutionCommon Probability Distibution
Common Probability DistibutionLukas Tencer
358 views22 Folien
Large Scale Online Learning of Image Similarity Through Ranking von
Large Scale Online Learning of Image Similarity Through RankingLarge Scale Online Learning of Image Similarity Through Ranking
Large Scale Online Learning of Image Similarity Through RankingLukas Tencer
588 views19 Folien
Slovakia Presentation at Day of Cultures von
Slovakia Presentation at Day of CulturesSlovakia Presentation at Day of Cultures
Slovakia Presentation at Day of CulturesLukas Tencer
305 views1 Folie
Web-based framework for online sketch-based image retrieval von
Web-based framework for online sketch-based image retrievalWeb-based framework for online sketch-based image retrieval
Web-based framework for online sketch-based image retrievalLukas Tencer
675 views10 Folien

Más de Lukas Tencer(12)

ICRA: Intelligent Platform for Collaboration and Interaction von Lukas Tencer
ICRA: Intelligent Platform for Collaboration and InteractionICRA: Intelligent Platform for Collaboration and Interaction
ICRA: Intelligent Platform for Collaboration and Interaction
Lukas Tencer630 views
Introduction to Probability von Lukas Tencer
Introduction to ProbabilityIntroduction to Probability
Introduction to Probability
Lukas Tencer651 views
Common Probability Distibution von Lukas Tencer
Common Probability DistibutionCommon Probability Distibution
Common Probability Distibution
Lukas Tencer358 views
Large Scale Online Learning of Image Similarity Through Ranking von Lukas Tencer
Large Scale Online Learning of Image Similarity Through RankingLarge Scale Online Learning of Image Similarity Through Ranking
Large Scale Online Learning of Image Similarity Through Ranking
Lukas Tencer588 views
Slovakia Presentation at Day of Cultures von Lukas Tencer
Slovakia Presentation at Day of CulturesSlovakia Presentation at Day of Cultures
Slovakia Presentation at Day of Cultures
Lukas Tencer305 views
Web-based framework for online sketch-based image retrieval von Lukas Tencer
Web-based framework for online sketch-based image retrievalWeb-based framework for online sketch-based image retrieval
Web-based framework for online sketch-based image retrieval
Lukas Tencer675 views
Supervised Learning of Semantic Classes for Image Annotation and Retrieval von Lukas Tencer
Supervised Learning of Semantic Classes for Image Annotation and RetrievalSupervised Learning of Semantic Classes for Image Annotation and Retrieval
Supervised Learning of Semantic Classes for Image Annotation and Retrieval
Lukas Tencer1.1K views
Personal Career,Education and skills presentation, 2011 von Lukas Tencer
Personal Career,Education and skills presentation, 2011Personal Career,Education and skills presentation, 2011
Personal Career,Education and skills presentation, 2011
Lukas Tencer502 views
Introduction to Computer Graphics, lesson 1 von Lukas Tencer
Introduction to Computer Graphics, lesson 1Introduction to Computer Graphics, lesson 1
Introduction to Computer Graphics, lesson 1
Lukas Tencer499 views
Computer graphics on web and in mobile devices von Lukas Tencer
Computer graphics on web and in mobile devicesComputer graphics on web and in mobile devices
Computer graphics on web and in mobile devices
Lukas Tencer423 views
Tracking of objects with known color signature - ELITECH 20 von Lukas Tencer
Tracking of objects with known color signature - ELITECH 20Tracking of objects with known color signature - ELITECH 20
Tracking of objects with known color signature - ELITECH 20
Lukas Tencer790 views

Último

Info Session November 2023.pdf von
Info Session November 2023.pdfInfo Session November 2023.pdf
Info Session November 2023.pdfAleksandraKoprivica4
11 views15 Folien
Top 10 Strategic Technologies in 2024: AI and Automation von
Top 10 Strategic Technologies in 2024: AI and AutomationTop 10 Strategic Technologies in 2024: AI and Automation
Top 10 Strategic Technologies in 2024: AI and AutomationAutomationEdge Technologies
18 views14 Folien
Report 2030 Digital Decade von
Report 2030 Digital DecadeReport 2030 Digital Decade
Report 2030 Digital DecadeMassimo Talia
15 views41 Folien
Melek BEN MAHMOUD.pdf von
Melek BEN MAHMOUD.pdfMelek BEN MAHMOUD.pdf
Melek BEN MAHMOUD.pdfMelekBenMahmoud
14 views1 Folie
Business Analyst Series 2023 - Week 3 Session 5 von
Business Analyst Series 2023 -  Week 3 Session 5Business Analyst Series 2023 -  Week 3 Session 5
Business Analyst Series 2023 - Week 3 Session 5DianaGray10
237 views20 Folien
virtual reality.pptx von
virtual reality.pptxvirtual reality.pptx
virtual reality.pptxG036GaikwadSnehal
11 views15 Folien

Último(20)

Business Analyst Series 2023 - Week 3 Session 5 von DianaGray10
Business Analyst Series 2023 -  Week 3 Session 5Business Analyst Series 2023 -  Week 3 Session 5
Business Analyst Series 2023 - Week 3 Session 5
DianaGray10237 views
From chaos to control: Managing migrations and Microsoft 365 with ShareGate! von sammart93
From chaos to control: Managing migrations and Microsoft 365 with ShareGate!From chaos to control: Managing migrations and Microsoft 365 with ShareGate!
From chaos to control: Managing migrations and Microsoft 365 with ShareGate!
sammart939 views
Special_edition_innovator_2023.pdf von WillDavies22
Special_edition_innovator_2023.pdfSpecial_edition_innovator_2023.pdf
Special_edition_innovator_2023.pdf
WillDavies2217 views
【USB韌體設計課程】精選講義節錄-USB的列舉過程_艾鍗學院 von IttrainingIttraining
【USB韌體設計課程】精選講義節錄-USB的列舉過程_艾鍗學院【USB韌體設計課程】精選講義節錄-USB的列舉過程_艾鍗學院
【USB韌體設計課程】精選講義節錄-USB的列舉過程_艾鍗學院
Data-centric AI and the convergence of data and model engineering: opportunit... von Paolo Missier
Data-centric AI and the convergence of data and model engineering:opportunit...Data-centric AI and the convergence of data and model engineering:opportunit...
Data-centric AI and the convergence of data and model engineering: opportunit...
Paolo Missier39 views
Attacking IoT Devices from a Web Perspective - Linux Day von Simone Onofri
Attacking IoT Devices from a Web Perspective - Linux Day Attacking IoT Devices from a Web Perspective - Linux Day
Attacking IoT Devices from a Web Perspective - Linux Day
Simone Onofri15 views
iSAQB Software Architecture Gathering 2023: How Process Orchestration Increas... von Bernd Ruecker
iSAQB Software Architecture Gathering 2023: How Process Orchestration Increas...iSAQB Software Architecture Gathering 2023: How Process Orchestration Increas...
iSAQB Software Architecture Gathering 2023: How Process Orchestration Increas...
Bernd Ruecker33 views
Piloting & Scaling Successfully With Microsoft Viva von Richard Harbridge
Piloting & Scaling Successfully With Microsoft VivaPiloting & Scaling Successfully With Microsoft Viva
Piloting & Scaling Successfully With Microsoft Viva
DALI Basics Course 2023 von Ivory Egg
DALI Basics Course  2023DALI Basics Course  2023
DALI Basics Course 2023
Ivory Egg16 views
Transcript: The Details of Description Techniques tips and tangents on altern... von BookNet Canada
Transcript: The Details of Description Techniques tips and tangents on altern...Transcript: The Details of Description Techniques tips and tangents on altern...
Transcript: The Details of Description Techniques tips and tangents on altern...
BookNet Canada135 views
Perth MeetUp November 2023 von Michael Price
Perth MeetUp November 2023 Perth MeetUp November 2023
Perth MeetUp November 2023
Michael Price19 views
The details of description: Techniques, tips, and tangents on alternative tex... von BookNet Canada
The details of description: Techniques, tips, and tangents on alternative tex...The details of description: Techniques, tips, and tangents on alternative tex...
The details of description: Techniques, tips, and tangents on alternative tex...
BookNet Canada126 views

Semi-Supervised Learning

  • 1. Semi-Supervised Learning Lukas Tencer PhD student @ ETS
  • 3. Image Similarity - Domain of origin :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 4. Face Recognition - Cross-race effect :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 5. Motivation in Machine Learning :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 6. Motivation in Machine Learning :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 8. When to use Semi-Supervised Learning? • Labelled data is hard to get and expensive – Speech analysis: • Switchboard dataset • 400 hours annotation time for 1 hour of speech – Natural Language Processing • Penn Chinese Treebank • 2 Years for 4000 sentences – Medical Application • Require experts opinion which might not be unique • Unlabelled data is cheap :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 9. Types of Semi-Supervised Leaning • Transductive Learning – Does not generalize to unseen data – Produces labels only for the data at training time • 1. Assume labels • 2. Train classifier on assumed labels • Inductive Learning – Does generalize to unseen data – Not only produces labels, but also the final classifier – Manifold Assumption :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 10. Selected Semi-Supervised Algorithms • Self-Training • Help-Training • Transductive SVM (S3VM) • Multiview Algorithms • Graph-Based Algorithms • Generative Models • ……. ….. … :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 11. Self-Training • The Idea: If I am highly confident in a label of examples, I am right • Given Training set 푇 = {푥푖 }, and unlabelled set 푈 = {푢푗 } 1. Train 푓 on 푇 2. Get predictions 푃 = 푓(푈) 3. If 푃푖 > 훼 then add (푥, 푓(푥)) to 푇 4. Retrain 푓 on 푇 :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 12. Self-Training • Advantages: – Very simple and fast method – Frequently used in NLP • Disadvantages: – Amplifies noise in labeled data – Requires explicit definition of 푃 푦 푥 – Hard to implement for discriminative classifiers (SVM) :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 13. Self-Training 1. Naïve Bayes Classifier on Bag-of-Visual-Word for 2 Classes 2. Classify Unlabelled Data base on Learned Classifier :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 14. Self-Training 3. Add the most confident images to the training set 4. Retrain and repeat :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 15. Help-Training • The Challenge: How to make Self-Training work for Discriminative Classifiers (SVM) ? • The Idea: Train Generative Help Classifier to get 푝(푦|푥) • Given Training set 푇 = {푥푖 }, unlabelled set 푈 = {푢푗 }, and generative classifier 푔 and discriminative classifier 푓 1. Train 푓 and 푔 on 푇 2. Get predictions 푃푔 = 푔(푈) and 푃푓 = 푓(푈) 3. If 푃푔,푖 > 훼 then add (푥, 푓(푥)) to 푇 4. Reduce the value of 훼 if |푃푔,푖 > 훼| = 0 5. Retrain 푓 and 푔 on 푇 until 푈 = 0 :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 16. Transductive SVM (S3VM) • The Idea: Find largest margin classifier, such that, unlabelled data are outside of the margin as much as possible, use regularization over unlabelled data • Given Training set 푇 = {푥푖 }, and unlabelled set 푈 = {푢푗 } 1. Find all possible labelings 푈1 ⋯ 푈푛 on 푈 2. For each 푇 푘 = 푇 ∪ 푈푘 train a standard SVM 3. Choose SVM with largest margins • What is the catch? • NP hard problem, fortunately approximations exist :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 17. Transductive SVM (S3VM) • Solving non-convex optimization problem: 퐽 휃 = • Methods: 1 2 푤 2 + 푐1 푥푖∈푇 퐿(푦푖푓휃 (푥푖 )) + 푐2 – Local Combinatorial Search – Standard unconstrained optimization solvers (CG, BFGS…) – Continuation Methods – Concave-Convex procedure (CCCP) – Branch and Bound :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data :: 푥푖∈푈 퐿( 푓휃 (푥푖 ) )
  • 18. Transductive SVM (S3VM) • Advantages: – Can be used with any SVM – Clear optimization criterion, mathematically well formulated • Disadvantages: – Hard to optimize – Prone to local minima – non convex – Only small gain given modest assumptions :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 19. Multiview Algorithms • The Idea: Train 2 classifiers on 2 disjoint sets of features, then let each classifier label unlabelled examples and teach the other classifier • Given Training set 푇 = {푥푖 }, and unlabelled set 푈 = {푢푗 } 1. Split 푇 into 푇1 and 푇2 on the feature dimension 2. Train 푓1 on 푇1 and 푓1 on 푇2 3. Get predictions 푃1 = 푓1(푈) and 푃2 = 푓2(푈) 4. Add: top 푘 from 푃1 to 푇2; top 푘 from 푃1 to 푇1 5. Repeat until 푈 = 0 :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 20. Multiview Algorithms • Application: Web-page Topic Classification – 1. Classifier for Images; 2. Classifier for Text :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 21. Multiview Algorithms • Advantages: – Simple Method applicable to any classifier – Can correct mistakes in classification between the 2 classifiers • Disadvantages: – Assumes conditional independence between features – Natural split may not exist – Artificial split may be complicated if only few eatures :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 22. Graph-Based Algorithms • The Idea: Create a connected graph from labelled and unlabelled examples, propagate labels over the graph :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 23. Graph-Based Algorithms • Advantages: – Great performance if graph fits the tasks – Can be used in combination with any model – Explicit mathematical formulation • Disadvantages: – Problem if graph does not fit the task – Hard to construct graph in sparse spaces :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 24. Generative Models • The Idea: Assume distribution using labelled data, update using unlabelled data • Simple models is: GMM + EM :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 25. Generative Models • Advantages: – Nice probabilistic framework – Instead of EM you can go full Bayesian and include prior with MAP • Disadvantages: – EM find only local minima – Makes strong assumptions about class distributions :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 26. What could go wrong? • Semi-Supervised Learning make a lot of assumptions – Smoothness – Clusters – Manifolds • Some techniques (Co-Training) require very specific setup • Frequently problem with noisy labels • There is no free lunch :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 27. There is much more out there • Structural Learning • Co-EM • Tri-Training • Co-Boosting • Unsupervised pretraining – deep learning • Transductive Inference • Universum Learning • Active Learning + Semi-Supervised Learning • ……. • ….. • … :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data :: My work
  • 28. Demo
  • 29. Conclusion • Play with Semi-Supervised Learning • Basic methods are vary simple to implement and can give you up to 5 to 10% accuracy • You can cheat at competitions by using unlabelled data, often no assumption is made about external data • Be careful when running Semi-Supervised Learning in production environment, keep an eye on your algorithm • If running in production, be aware that data patterns change and old assumptions about labels may screw up you new unlabelled data :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 30. Some more resources Videos to watch: Semisupervised Learning Approaches – Tom Mitchell CMU : http://videolectures.net/mlas06_mitchell_sla/ MLSS 2012 Graph based semi-supervised learning - Zoubin Ghahramani Cambridge : https://www.youtube.com/watch?v=HZQOvm0fkLA Books to read: • Semi-Supervised Learning – Chapelle, Schölkopf, Zien • Introduction to Semi-Supervised Learning - Zhu, Oldberg, Brachman, Dietterich :: Semi-Supervised Learning :: Lukas Tencer :: MTL Data ::
  • 31. THANKS FOR YOUR TIME Lukas Tencer lukas.tencer@gmail.com http://lukastencer.github.io/ https://github.com/lukastencer https://twitter.com/lukastencer Graduating August 2015, looking for ML and DS opportunities