Semi-Supervised Learning over Text

Tom M. MitchellMachine Learning Department

Carnegie Mellon University

September 2006

Statistical learning methods require LOTS of training data

Can we use all that unlabelled text?

Outline

• Maximizing likelihood in probabilistic models– EM for text classification

• Co-Training and redundantly predictive features– Document classification– Named entity recognition– Theoretical analysis

• Sample of additional tasks– Word sense disambiguation– Learning HTML-based extractors– Large-scale bootstrapping: extracting from the web

Many text learning tasks

• Document classification. – f: Doc Class

– Spam filtering, relevance rating, web page classification, ...

– and unsupervised document clustering

• Information extraction. – f: Sentence Fact, f: Doc Facts

• Parsing– f: Sentence ParseTree

– Related: part-of-speech tagging, co-reference res., prep phrase attachment

• Translation– f: EnglishDoc FrenchDoc

1. Semi-supervised Document classification (probabilistic model and EM)

Document Classification: Bag of Words Approach

aardvark 0

about 2

all 2

Africa 1

apple 0

anxious 0

...

gas 1

...

oil 1

…

Zaire 0

For code and data, see www.cs.cmu.edu/~tom/mlbook.html click on “Software and Data”

Accuracy vs. # training examples

What if we have labels for only some documents?

Y

X1 X4X3X2

Y X1 X2 X3 X4

1 0 0 1 1

0 0 1 0 0

0 0 0 1 0

? 0 1 1 0

? 0 1 0 1

Learn P(Y|X)

EM: Repeat until convergence

1. Use probabilistic labels to train classifier h

2. Apply h to assign probabilistic labels to unlabeled data

From [Nigam et al., 2000]

E Step:

M Step:wt is t-th word in vocabulary

Using one labeled example per class

Words sorted by P(w|course) / P(w| : course)

20 Newsgroups

20 Newsgroups

Elaboration 1: Downweight the influence of unlabeled

examples by factor

New M step:Chosen by cross validation

Why/When will this work?

• What’s best case? Worst case? How can we test which we have?

EM for Semi-Supervised Doc Classification

• If all data is labeled, corresponds to supervised training of Naïve Bayes classifier

• If all data unlabeled, corresponds to mixture-of-multinomial clustering

• If both labeled and unlabeled data, it helps if and only if the mixture-of-multinomial modeling assumption is correct

• Of course we could extend this to Bayes net models other than Naïve Bayes (e.g., TAN tree)

• Other extensions: model negative class as mixture of N multinomials

2. Using Redundantly Predictive Features (Co-Training)

Redundantly Predictive Features

Professor Faloutsos my advisor

Co-Training

Answer1

Classifier1

Answer2

Classifier2

Key idea: Classifier1 and ClassifierJ must:

1. Correctly classify labeled examples

2. Agree on classification of unlabeled

CoTraining Algorithm #1 [Blum&Mitchell, 1998]

Given: labeled data L,

unlabeled data U

Loop:

Train g1 (hyperlink classifier) using L

Train g2 (page classifier) using L

Allow g1 to label p positive, n negative examps from U

Allow g2 to label p positive, n negative examps from U

Add these self-labeled examples to L

CoTraining: Experimental Results

• begin with 12 labeled web pages (academic course)

• provide 1,000 additional unlabeled web pages

• average error: learning from labeled data 11.1%;

• average error: cotraining 5.0%

Typical run:

Co-Training for Named Entity Extraction(i.e.,classifying which strings refer to people, places, dates, etc.)

Answer1

Classifier1

Answer2

Classifier2

I flew to New York today.

New York I flew to ____ today

[Riloff&Jones 98; Collins et al., 98; Jones 05]

One result [Blum&Mitchell 1998]: • If

– X1 and X2 are conditionally independent given Y– f is PAC learnable from noisy labeled data

• Then– f is PAC learnable from weak initial classifier plus unlabeled

data

CoTraining setting:

• wish to learn f: X Y, given L and U drawn from P(X)

• features describing X can be partitioned (X = X1 x X2)

such that f can be computed from either X1 or X2

Co-Training Rote Learner

My advisor+

-

-

pageshyperlinks

-

--

-

+

+

-

-

-

+

++

+

-

-

+

+

-

Co Training

• What’s the best-case graph? (most benefit from unlabeled data)

• What the worst case?• What does conditional-independence imply about

graph?

x1 x2

+

-

+

Expected Rote CoTraining error given m examples

mjj

j

gxPgxPerrorE ))(1)(( Where g is the jth connected component of graph of L+U, m is number of labeled examples

j

)()()()(,

:

:

221121

21

xfxgxgxggand

ondistributiunknownfromdrawnxwhere

XXXwhere

YXflearn

settingCoTraining

How many unlabeled examples suffice?

Want to assure that connected components in the underlying distribution, GD, are connected components in the observed sample, GS

GD GS

O(log(N)/) examples assure that with high probability, GS has same connected components as GD [Karger, 94]

N is size of GD, is min cut over all connected components of GD

PAC Generalization Bounds on CoTraining

[Dasgupta et al., NIPS 2001]

This theorem assumes X1 and X2 are conditionally independent given Y

Co-Training Theory

Final Accuracy

# unlabeled examples

dependencies among input features

# Redundantly predictive inputs

# labeled examples

Correctness of confidence assessments

How can we tune learning environment to enhance effectiveness of Co-Training?

best: inputs conditionally indep given class, increased number of redundant inputs, …

• Idea: Want classifiers that produce a maximally consistent labeling of the data

• If learning is an optimization problem, what function should we optimize?

What if CoTraining Assumption Not Perfectly Satisfied?

-

+

+

+

Example 2: Learning to extract named entities

I arrived in Beijing on Saturday.

location?

If: “I arrived in <X> on Saturday.”

Then: Location(X)

Co-Training for Named Entity Extraction(i.e.,classifying which strings refer to people, places, dates, etc.)

Answer1

Classifier1

Answer2

Classifier2

I arrived in Beijing saturday.

Beijing I arrived in __ saturday

[Riloff&Jones 98; Collins et al., 98; Jones 05]

Bootstrap learning to extract named entities[Riloff and Jones, 1999], [Collins and Singer, 1999], ...

Iterations

InitializationAustraliaCanada China England France Germany Japan Mexico Switzerland United_states

locations in ?x

South AfricaUnited KingdomWarrentonFar_EastOregonLexingtonEuropeU.S._A.Eastern CanadaBlairSouthwestern_statesTexasStatesSingapore …

operations in ?x

ThailandMaineproduction_controlnorthern_LosNew_Zealandeastern_EuropeAmericasMichigan New_HampshireHungarysouth_americadistrictLatin_AmericaFlorida ...

republic of ?x

…...

Co-EM [Nigam & Ghani, 2000; Jones 2005]

Idea:• Like co-training, use one set of features to label the other• Like EM, iterate

– Assigning probabilistic values to unobserved class labels– Updating model parameters (= labels of other feature set)

CoEM applied to Named Entity Recognition[Rosie Jones, 2005], [Ghani & Nigam, 2000]

Update rules:

CoEM applied to Named Entity Recognition[Rosie Jones, 2005], [Ghani & Nigam, 2000]

Update rules:

CoEM applied to Named Entity Recognition[Rosie Jones, 2005], [Ghani & Nigam, 2000]

Update rules:

[Jones, 2005]

Can use this for active learning...

[Jones, 2005]

• Idea: Want classifiers that produce a maximally consistent labeling of the data

• If learning is an optimization problem, what function should we optimize?

What if CoTraining Assumption Not Perfectly Satisfied?

-

+

+

+

What Objective Function?

2

2211

,

22211

,

222

,

211

43

2

)(ˆ)(ˆ

||||

1

||

14

))(ˆ)(ˆ(3

))(ˆ(2

))(ˆ(1

4321

ULxLyx

Ux

Lyx

Lyx

xgxg

ULy

LE

xgxgE

xgyE

xgyE

EcEcEEE

Error on labeled examples

Disagreement over unlabeled

Misfit to estimated class priors

What Function Approximators?

• Same functional form as logistic regression

• Use gradient descent to simultaneously learn g1 and g2, directly minimizing E = E1 + E2 + E3 + E4

• No word independence assumption, use both labeled and unlabeled data

j

jj xw

exg

1,

1

1)(ˆ1

jjj xw

exg

2,

1

1)(ˆ2

Gradient CoTraining Classifying Capitalized sequences as Person Names

25 labeled 5000 unlabeled

2300 labeled 5000 unlabeled

Using labeled data only

Cotraining

Cotraining without fitting class priors (E4)

.27

.13.24

* sensitive to weights of error terms E3 and E4

.11 *.15 *

*

Error Rates

Eg., “Company president Mary Smith said today…”x1 x2 x1

Example 3: Word sense disambiguation [Yarowsky]

• “bank” = river bank, or financial bank??

• Assumes a single word sense per document– X1: the document containing the word– X2: the immediate context of the word (‘swim near the __’)

Successfully learns “context word sense” rules when word occurs multiples times in documents.

Example 4: Bootstrap learning for IE from HTML structure

[Muslea, et al. 2001]

X1: HTML preceding the target

X2: HTML following the target

Example Bootstrap learning algorithms:

• Classifying web pages [Blum&Mitchell 98; Slattery 99]

• Classifying email [Kiritchenko&Matwin 01; Chan et al. 04]

• Named entity extraction [Collins&Singer 99; Jones&Riloff 99]

• Wrapper induction [Muslea et al., 01; Mohapatra et al. 04]

• Word sense disambiguation [Yarowsky 96]

• Discovering new word senses [Pantel&Lin 02]

• Synonym discovery [Lin et al., 03]

• Relation extraction [Brin et al.; Yangarber et al. 00]

• Statistical parsing [Sarkar 01]

What to Know

• Several approaches to semi-supervised learning– EM with probabilistic model– Co-Training– Graph similarity methods– ...– See reading list below

• Redundancy is important• Much more to be done:

– Better theoretical models of when/how unlabeled data can help– Bootstrap learning from the web (e.g. Etzioni, 2005, 2006)– Active learning (use limited labeling time of human wisely)– Never ending bootstrap learning?– ...

Further Reading

• Semi-Supervised Learning, Olivier Chapelle, Bernhard Sch¨olkopf, and Alexander Zien (eds.), MIT Press, 2006.

• Semi-Supervised Learning Literature Survey, Xiaojin Zhu, 2006.

• Unsupervised word sense disambiguation rivaling supervised methods D. Yarowsky (1995)

• "Semi-Supervised Text Classification Using EM,"  K. Nigam, A. McCallum, and T. Mitchell, in Semi-Supervised Learning, Olivier Chapelle, Bernhard Sch¨olkopf, and Alexander Zien (eds.), MIT Press, 2006.

• " Text Classification from Labeled and Unlabeled Documents using EM," K. Nigam, Andrew McCallum, Sebastian Thrun and Tom Mitchell. Machine Learning, Kluwer Academic Press, 1999.

• " Combining Labeled and Unlabeled Data with Co-Training," A. Blum and T. Mitchell, Proceedings of the 1998 Conference on Computational Learning Theory, July 1998.

• Discovering Word Senses from Text Pantel & Lin (2002) • Creating Subjective and Objective Sentence Classifiers from Unannotated Texts by

Janyce Wiebe and Ellen Riloff (2005) • Graph Based Semi-Supervised Approach for Information Extraction by Hany Hassan,

Ahmed Hassan and Sara Noeman (2006) • The use of unlabeled data to improve supervised learning for text summarization by

MR Amini, P Gallinari (2002)

Further Reading

• Yusuke Shinyama and Satoshi Sekine. Preemptive Information Extraction using Unrestricted Relation Discovery

• Alexandre Klementiev and Dan Roth. Named Entity Transliteration and Discovery from Multilingual Comparable Corpora.

• Rion L. Snow, Daniel Jurafsky, Andrew Y. Ng. Learning syntactic patterns for automatic hypernym discovery

• Sarkar. (1999). Applying Co-training methods to Statistical Parsing. • S. Brin, 1998. Extracting patterns and relations from the World Wide Web,  EDBT'98 • O. Etzioni et al., 2005. "Unsupervised Named-Entity Extraction from the Web: An

Experimental Study," AI Journal, 2005.