INFORMATION EXTRACTIONNaïve Bayes: Mixture of Multinomials Model 1. Pick the class: P(class) 2. For...

1of 63

INFORMATION EXTRACTION

CS8691-AI-AI-UNITV-INFORMATION EXTRACTION

2of 63

Text Classification by Example


3of 63



4of 63



5of 63



6of 63



7of 63

How could you build a text classifier?

• Take some ideas from machine learning– Supervised learning setting– Examples of each class (a few or thousands)

• Take some ideas from machine translation– Generative models– Language models

• Simplify each and stir thoroughly


8of 63

Basic Approach of Generative Modeling

1. Pick representation for data

2. Write down probabilistic generative model

3. Estimate model parameters with training data

4. Turn model around to calculate unknown values for new data


9 of 63

Naïve Bayes: Bag of Words Representation

Corn prices rose today while corn futures dropped in surprising trading activity. Corn ...

activity 1 cable 0 corn 3 damp 0

drawer 0 dropped 1 elbow 0

earning 0 . . . . . .

All words in dictionary

Occurrence counts


10of 63

Naïve Bayes: Mixture of Multinomials Model1. Pick the class: P(class)

2. For every word, pick from the class urn: P(word|class)

while

polo

socceractivity

droppedsoccer

the

ball

COMPUTERS SPORTS

thein

web

windows

the

thein

java

windows

again

modem

Word independence assumption!CS8691-AI-AI-UNITV-INFORMATION EXTRACTION

11of 63

Naïve Bayes: Estimating Parameters• Just like estimating biased coin flip probabilities

• Estimate MAP word probabilities:

• Estimate MAP class priors:

classdoc

classdoc

docNVocab

docwordclassword

)(||

),N(1)|P(

)N()N(),N(1)P(

docclassclassdocclass


12of 63

Naïve Bayes: Performing Classification

• Word independence assumption

• Take the class with the highest probability

docword

classwordclass )|P()P(

)(P)(P)|(P)|(P

docclassclassdocdocclass


13of 63

Classification Tricks of the Trade• Stemming

– run, runs, running, ran run– table, tables, tabled table– computer, compute, computing compute

• Stopwords– Very frequent function words generally uninformative– if, in, the, like, …

• Information gain feature selection– Keep just most indicative words in the vocabulary


14of 63

Naïve Bayes Rules of Thumb

• Need hundreds of labeled examples per class for good performance (~85% accuracy)

• Stemming and stopwords may or may not help

• Feature selection may or may not help

• Predicted probabilities will be very extreme

• Use sum of logs instead of multiplying probabilities for underflow prevention

• Coding this up is trivial, either as a mapreduce or not


15of 63

Information Extraction with Generative Models


16of 63

Example: A Problem

Genomics job

Mt. Baker, the school district

Baker Hostetler, the company

Baker, a job opening


17of 63

Example: A Solution


18of 63

Job Openings:Category = Food ServicesKeyword = BakerLocation = Continental U.S.


19 of 63

Extracting Job Openings from the Web

Title: Ice Cream Guru

Description: If you dream of cold creamy…

Contact:[email protected]

Category:Travel/Hospitality

Function:Food Services


20of 63

Potential Enabler of Faceted Search


21of 63

Lots of Structured Information in Text


22of 63

IE from Research Papers


23of 63

What is Information Extraction?

• Recovering structured data from formatted text


24of 63


• Recovering structured data from formatted text– Identifying fields (e.g. named entity recognition)


25of 63



– Understanding relations between fields (e.g. record association)


26of 63




– Normalization and deduplication


27of 63




– Normalization and deduplication

• Today, focus on field identification


28of 63

IE HistoryPre-Web• Mostly news articles

– De Jong’s FRUMP [1982]• Hand-built system to fill Schank-style “scripts” from news wire

– Message Understanding Conference (MUC) DARPA [’87-’95], TIPSTER [’92-’96]

• Most early work dominated by hand-built models– E.g. SRI’s FASTUS, hand-built FSMs.– But by 1990’s, some machine learning: Lehnert, Cardie, Grishman and then HMMs: Elkan

[Leek ’97], BBN [Bikel et al ’98]

Web• AAAI ’94 Spring Symposium on “Software Agents”

– Much discussion of ML applied to Web. Maes, Mitchell, Etzioni.

• Tom Mitchell’s WebKB, ‘96– Build KB’s from the Web.

• Wrapper Induction– Initially hand-build, then ML: [Soderland ’96], [Kushmeric ’97],…


29 of 63

IE Posed as a Machine Learning Task

• Training data: documents marked up with ground truth

• In contrast to text classification, local features crucial. Features of:– Contents

– Text just before item

– Text just after item

– Begin/end boundaries

00 : pm Place : Wean Hall Rm 5409 Speaker : Sebastian Thrun

prefix contents suffix

… …


30of 63

Good Features for Information Extraction

Example word features:– identity of word

– is in all caps

– ends in “-ski”

– is part of a noun phrase

– is in a list of city names

– is under node X in WordNet or Cyc

– is in bold font

– is in hyperlink anchor

– features of past & future

– last person name was female

– next two words are “and Associates”

begins-with-number

begins-with-ordinal

begins-with-punctuation

begins-with-question-word

begins-with-subject

blank

contains-alphanum

contains-bracketed-number

contains-http

contains-non-space

contains-number

contains-pipe

contains-question-mark

contains-question-word

ends-with-question-mark

first-alpha-is-capitalized

indented

indented-1-to-4

indented-5-to-10

more-than-one-third-space

only-punctuation

prev-is-blank

prev-begins-with-ordinal

shorter-than-30

Creativity and Domain Knowledge Required!


31of 63

Is Capitalized

Is Mixed Caps

Is All Caps

Initial Cap

Contains Digit

All lowercase

Is Initial

Punctuation

Period

Comma

Apostrophe

Dash

Preceded by HTML tag

Character n-gram classifier says string is a person name (80% accurate)

In stopword list(the, of, their, etc)

In honorific list(Mr, Mrs, Dr, Sen, etc)

In person suffix list(Jr, Sr, PhD, etc)

In name particle list (de, la, van, der, etc)

In Census lastname list;segmented by P(name)

In Census firstname list;segmented by P(name)

In locations lists(states, cities, countries)

In company name list(“J. C. Penny”)

In list of company suffixes(Inc, & Associates, Foundation)

Word Features– lists of job titles, – Lists of prefixes– Lists of suffixes– 350 informative phrases

HTML/Formatting Features– {begin, end, in} x

{, , <a>, <hN>} x{lengths 1, 2, 3, 4, or longer}

– {begin, end} of line

Creativity and Domain Knowledge Required!Good Features for Information Extraction

32of 63

Landscape of ML Techniques for IE:

Any of these models can be used to capture words, formatting or both.

Classify Candidates

Abraham Lincolnwas born in Kentucky.

Classifier

which class?

Sliding Window

Abraham Lincoln was born in Kentucky.

Classifier

which class?

Try alternatewindow sizes:

Boundary Models


Classifier

which class?

BEGIN END BEGIN END

BEGIN

Finite State Machines


Most likely state sequence?

Wrapper Induction

Abraham Lincoln was born in Kentucky.

Learn and apply pattern for a website





PersonName

33of 63

Sliding Windows & Boundary Detection


34of 63

Information Extraction by Sliding WindowsGRAND CHALLENGES FOR MACHINE LEARNING

Jaime CarbonellSchool of Computer ScienceCarnegie Mellon University

3:30 pm7500 Wean Hall

Machine learning has evolved from obscurity in the 1970s into a vibrant and popular discipline in artificial intelligence during the 1980s and 1990s. As a result of its success and growth, machine learning is evolving into a collection of related disciplines: inductive concept acquisition, analytic learning in problem solving (e.g. analogy, explanation-based learning), learning theory (e.g. PAC learning), genetic algorithms, connectionist learning, hybrid systems, and so on.

CMU UseNet Seminar Announcement

E.g.Looking forseminarlocation


35of 63

Information Extraction by Sliding WindowsGRAND CHALLENGES FOR MACHINE LEARNING







36of 63

Information Extraction by Sliding Window

GRAND CHALLENGES FOR MACHINE LEARNING







37of 63









38of 63









39 of 63

Information Extraction with Sliding Windows[Freitag 97, 98; Soderland 97; Califf 98]

00 : pm Place : Wean Hall Rm 5409 Speaker : Sebastian Thrunw t-m w t-1 w t w t+n w t+n+1 w t+n+m

prefix contents suffix

… …

• Standard supervised learning setting– Positive instances: Windows with real label

– Negative instances: All other windows

– Features based on candidate, prefix and suffix

• Special-purpose rule learning systems work wellcourseNumber(X) :-

tokenLength(X,=,2), every(X, inTitle, false), some(X, A, <previousToken>, inTitle, true),some(X, B, <>. tripleton, true)CS8691-AI-AI-UNITV-INFORMATION EXTRACTION

40of 63

IE by Boundary DetectionGRAND CHALLENGES FOR MACHINE LEARNING







41of 63








42of 63








43of 63








44of 63








45of 63

BWI: Learning to detect boundaries

• Another formulation: learn three probabilistic classifiers:– START(i) = Prob( position i starts a field)– END(j) = Prob( position j ends a field)– LEN(k) = Prob( an extracted field has length k)

• Then score a possible extraction (i,j) bySTART(i) * END(j) * LEN(j-i)

• LEN(k) is estimated from a histogram

• START(i) and END(j) learned by boosting over simple boundary patterns and features

[Freitag & Kushmerick, AAAI 2000]


46of 63

Problems with Sliding Windows and Boundary Finders

• Decisions in neighboring parts of the input are made independently from each other.

– Sliding Window may predict a “seminar end time” before the “seminar start time”.

– It is possible for two overlapping windows to both be above threshold.

– In a Boundary-Finding system, left boundaries are laid down independently from right boundaries, and their pairing happens as a separate step.


47of 63

Hidden Markov Models


48of 63

Citation Parsing

• Fahlman, Scott & Lebiere, Christian(1989). The cascade-correlation learning architecture.Advances in Neural Information Processing Systems,pp. 524-532.

• Fahlman, S.E. and Lebiere, C., “The Cascade Correlation Learning Architecture,”Neural Information Processing Systems,pp. 524-532, 1990.

• Fahlman, S. E.(1991)The recurrent cascade-correlation learning architecture. NIPS 3, 190-205.


49 of 63

Can we do this with probabilistic generative models?

• Could have classes for {author, title, journal, year, pages}

• Could classify every word or sequence?– Which sequences?

• Something interesting in the sequence of fields that we’d like to capture– Authors come first– Title comes before journal– Page numbers come near the end


50of 63

Hidden Markov Models: The Representation

• A document is a sequence of words

• Each word is tagged by its class

• fahlman s e and lebiere c the cascade correlation learning architectureneural information processing systemspp 524 532 1990


51of 63

HMM: Generative Model (1)

Author Title Journal

Year Pages


52of 63


Author Title

Year Pages


53of 63


• States: xi

• State transitions: P(xi|xj) = a[xi|xj] • Output probabilities: P(oi|xj) = b[oi|xj]

• Markov independence assumptionCS8691-AI-AI-UNITV-INFORMATION EXTRACTION

54of 63

HMMs: Estimating Parameters• With fully-labeled data, just like naïve Bayes

• Estimate MAP output probabilities:

• Estimate MAP state transitions:

j

j

xdataword

xdatawordi

ji Vocab

wordoxo

@

@

1||

),N(1]|[b

datax

dataxij

ji

j

j

x

xxxxa

1||

),N(1]|[

1


55of 63

HMMs: Performing Extraction

• Given output words:– fahlman s e 1991 the recurrent cascade correlation learning architecture nips 3 190 205

• Find state sequence that maximizes:

• Lots of possible state sequences to test (514)

Hmm…

i

iiii xobxxa ]|[]|[ 1


56of 63

Representation for Paths: Trellis


57of 63

Representation for Paths: Trellis


58of 63CS8691-AI-AI-UNITV-INFORMATION EXTRACTION

59 of 63CS8691-AI-AI-UNITV-INFORMATION EXTRACTION

60of 63CS8691-AI-AI-UNITV-INFORMATION EXTRACTION

61of 63

HMM Example: NymbleTask: Named Entity Extraction

Train on 450k words of news wire text.

Case Language F1 .Mixed English 93%Upper English 91%Mixed Spanish 90%

[Bikel, et al 97]

Person

Org

Other

(Five other name classes)

start-of-sentence

end-of-sentence

Results:

• Bigram within classes

• Backoff to unigram

• Special capitalization and number features…


62of 63

Nymble word features


63of 63

HMMs: A Plethora of Applications

• Information extraction• Part of speech tagging• Word segmentation

• Gene finding• Protein structure prediction

• Speech recognition

• Economics, Climatology, Robotics, …CS8691-AI-AI-UNITV-INFORMATION EXTRACTION

INFORMATION EXTRACTIONNaïve Bayes: Mixture of Multinomials Model 1. Pick the class: P(class) 2. For...

Documents

Transcript of INFORMATION EXTRACTIONNaïve Bayes: Mixture of Multinomials Model 1. Pick the class: P(class) 2. For...