Augmenting WordNet for Deep Understanding of Text

“Deep Understanding”• Not just parsing + word senses • Construction of a coherent representation of the scene the text describes • Challenge: much of that representation is not in the t

Trang 1

Augmenting WordNet for Deep

Understanding of Text

Peter Clark, Phil Harrison,

Bill Murray, John Thompson (Boeing)

Christiane Fellbaum (Princeton Univ)

Jerry Hobbs (ISI/USC)

Trang 2

“Deep Understanding”

• Not (just) parsing + word senses

• Construction of a coherent representation of the scene the text describes

• Challenge: much of that representation is not in the text

“A soldier

was killed in

a gun battle”

“The soldier died”

“The soldier was shot”

“There was a fight”

…

Trang 3

“A soldier

was killed in

a gun battle”

Guns can kill

If you are killed, you are dead

…

How do we get this

knowledge into the

machine?

How do we exploit it?

Trang 4

“A soldier

was killed in

a gun battle”

Guns can kill

If you are killed, you are dead

…

Several partially useful

resources exist.

WordNet is already used

a lot…can we extend it?

Trang 5

The Initial Vision

• Our vision:

Rapidly expand WordNet to be more of a knowledge-baseQuestion-answering software to demonstrate its use

Trang 6

The Evolution of WordNet

– introduce the instance/class distinction

• Paris isa Capital-City is-type-of City – add in some derivational links

• explode related-to explosion

• …

lexical

resource

knowledge

Trang 7

Augmenting WordNet

• World Knowledge

– Sense-disambiguate the glosses (by hand)

– Convert the glosses to logic

• Similar to LCC’s Extended WordNet attempt– Axiomatize “core theories”

• WordNet links

– Morphosemantic links

– Purpose links

• Experiments

Trang 8

Converting the Glosses to Logic

Convert gloss to form “word is gloss”

Parse (Charniak)

“ambition#n2: A strong drive for success”

LFToolkit: Generate logical form fragments

strong drive for success strong(x1) & drive(x2) & for(x3,x4) & success(x5)

Lexical output rules produce logical form

fragments

Trang 9

Parse (Charniak) LFToolkit: Generate logical form fragments

Identify equalities , add senses

Trang 10

A strong drive for success strong(x1) & drive(x2) & for(x3,x4) & success(x5)

x2=x3 x1=x2

Lexical output rules produce logical form

fragments

Composition rules identify variables

x4=x5

Trang 11

Parse (Charniak) LFToolkit: Generate logical form fragments

ambition#n2(x1) → a(x1) & strong#a1(x1) & drive#n2(x1) & for(x1,x2) & success#a3(x2)

Trang 12

• But often not Primary problems:

1 Errors in the language processing

2 Only capture definitional knowledge

3 “flowery” language, many gaps, metonymy, ambiguity;

If logic closely follows syntax → “logico-babble”

“hammer#n2: tool used to deliver an impulsive force by striking”

Trang 13

Augmenting WordNet

– Axiomatize “core theories”

– Purpose links

• Experiments

Trang 14

Core Theories

• Many domain-specific facts are instantiations of more general, “core” knowledge

• By encoding this core knowledge, get leverage

• eg 517 “vehicle” noun (senses), 185 “cover” verb (senses)

• Approach:

– Analysis and grouping of words in Core WordNet

– Identification and encoding of underlying theories

Trang 16

Augmenting WordNet

– Axiomatize “core theories”

– Purpose links

• Experiments

Trang 17

Morphosemantic Links

• Often need to cross part-of-speech

T: A council worker cleans up after Tuesday's violence in Budapest.H: There were attacks in Budapest on Tuesday

(“attack”) attack_v3 aggression_n4 (←“violence”)

“aggress”/“aggression”

derivation link

• Can solve with WN’s derivation links:

Trang 18

• But can go wrong!

T: Paying was slowH1: The transaction was slowH2: *The person was slow [NOT entailed]

“pay”/“payment”payment_n1 (→ “transaction”)

Trang 19

• Task: Classify the 22,000 links in WordNet:

• Semi-automatic process

– Exploit taxonomy and morphology

• 15 semantic types used

– agent, undergoer, instrument, result, material, destination, location, result, by-means-of, event, uses, state, property,

Verb Synset Noun Synset Relationship

hammer_v1 hammer_n1 instrumentexecute_v1 execution_n1 event (equal)

Trang 20

Experimentation

Trang 21

Task: Recognizing Entailment

• Experiment with WordNet, logical glosses, DIRT

• Text interpretation to logic using Boeing’s NLP system

• Entailment: T → H if:

– T is subsumed by H (“cat eats mouse” → “animal was eaten”)– An elaboration of T using inference rules is subsumed by H

• (“cat eats mouse” → “cat swallows mouse”)

“A soldier was

isa(soldier01,soldier_n1), isa(……

object(kill01,soldier01) during(kill01,battle01) instrument(battle01,gun01)

Trang 22

Successful Examples with the Glosses

• Good example

T: Britain puts curbs on immigrant labor from Bulgaria and Romania.H: Britain restricted workers from Bulgaria

14.H4

Trang 23

Trang 24

T: The administration managed to track down the perpetrators.

H: The perpetrators were being chased by the administration

56.H3

• Another (somewhat) good example

Trang 25

T: The administration managed to track down the perpetrators.

WN: hunt_v1 “hunt” “track down”: pursue for food or sport

56.H3

T: The administration managed to pursue the perpetrators [for food

or sport!]

• Another (somewhat) good example

Trang 26

Unsuccessful examples with the glosses

• More common: Being “tantalizingly close”

T: Satomi Mitarai bled to death

H: His blood flowed out of his body

16.H3

Trang 27

T: Satomi Mitarai bled to death

H: His blood flowed out of his body

Trang 28

T: The National Philharmonic orchestra draws large crowds.H: Large crowds were drawn to listen to the orchestra.

20.H2

Trang 29

T: The National Philharmonic orchestra draws large crowds.H: Large crowds were drawn to listen to the orchestra.

20.H2

WN: orchestra = collection of musicians WN: musician: plays musical instrument

WN: music = sound produced by musical instruments

WN: listen = hear = perceive sound

WordNet:

So close!

Trang 30

Success with Morphosemantic Links

• Good example

T: The Zoopraxiscope was invented by Mulbridge

H*: Mulbridge was the invention of the Zoopraxiscope [NOT entailed]

Trang 31

T: The president visited Iraq in September.

H: The president traveled to Iraq

54.H1

Successful Examples with DIRT

• Good example

DIRT: IF Y is visited by X THEN X flocks to Y

WordNet: "flock" is a type of "travel"

 Entailed [correct]

Trang 32

T: The US troops stayed in Iraq although the war was over.

H*: The US troops left Iraq when the war was over [NOT entailed]

Trang 33

Overall Results

• Note: Eschewing statistics!

• BPI test suite (61%):

Correct Incorrect

When H or ¬H is predicted by:

WordNet taxonomy + morphosemantics 14 1

When H or ¬H is not predicted: 97 72

“Straight-Forward”

Trang 34

Overall Results

Correct Incorrect

Useful

Trang 35

Overall Results

Correct Incorrect

Occasionally

useful

Trang 36

Overall Results

Correct Incorrect

Often useful

but

unreliable

• RTE3: 55%

Trang 38

Thank you!

Định dạng
Số trang	38
Dung lượng	0,91 MB