BURC: Bootstrapping Using ResearchCyc

1
BURC Bootstrapping Using ResearchCyc

• By Kino Coursey

2
Introduction to the Problem

• Goal To extend Cycs knowledge base using
  relationships implied to be possible, normal or
  commonplace in the world
• Prior work with Cyc knowledge entry has been
  manually oriented
• How will we collect commonsense without a body
  and manual labor?
• Read, Parse, Mine!
• Proposal Read text, Parse into a database,
  Extract relations between words, Propose
  hypothetical relations between concepts

3
Basic Analogy

• The Shotgun approach to the Human Genome
• Extract millions of fragments then knit them back
  together by finding commonalities
• Will it work for the Human Menome?

4
What is Cyc?

• the world's largest and most complete general
  knowledge base and commonsense reasoning engine
• Started in mid 1980s (should take only 10
  years.)
• Logic Based
• LISP oriented
• For WordNet users, each Concept Synset
• Available from http//www.opencyc.org
• http//researchcyc.cyc.com

• Big (ResearchCyc v0.8)
• Constants 89,379
• Assertions 968,985
• Deduction 361,185

• Sample Collection Extents
• EnglishWord 18,007
• Event 6,050
• PartiallyTangible 24,387
• Microtheory 1,688

5
Example of what Cyc currently knows about fingers

• Collection Finger
• GAF Arg 1
• Mt UniversalVocabularyMtisa
  AnimalBodyPartType
• quotedIsa DensoOntologyConstant
• genls Digit-AnatomicalPart
• comment "The collection of all digits of all
  Hands (q.v.). Fingers are (typically) flexibly
  jointed and are necessary to enabling the hand
  (and its owner) to perform grasping and
  manipulation actions."
• Mt BaseKBdefiningMt AnimalPhysiologyVocabula
  ryMt
• Mt AnimalPhysiologyMtproperPhysicalPartTypes
  Fingernail
• Mt WordNetMappingMt (synonymousExternalConcept
  Finger WordNet-Version2_0 "N05247839")
  (synonymousExternalConcept Finger
  WordNet-1997Version "N04312497")

• GAF Arg 2
• Mt UniversalVocabularyMt (genls LittleFinger
  Finger) (genls IndexFinger Finger) (genls Thumb
  Finger) (genls RingFinger Finger) (genls
  MiddleFinger Finger)
• Mt HumanActivitiesMt (bodyPartsUsed-TypeType
  Typing Finger)
• Mt HumanSocialLifeMt (bodyPartsUsed-TypeType
  PointingAFinger Finger)

6
Example of what Cyc currently knows about fingers
- 2

• GAF Arg 3
• Mt HumanPhysiologyMt (relationAllExists
  anatomicalParts HomoSapiens Finger)
• Mt VertebratePhysiologyMt (relationAllExistsCou
  nt physicalParts Hand Finger 5)
• Mt UniversalVocabularyMt (relationAllOnly
  wornOn Ring-Jewelry Finger)
• Mt AnimalPhysiologyMt (relationExistsAll
  physicalParts Hand Finger)
• GAF Arg 4
• Mt GeneralEnglishMt (denotation Finger-TheWord
  CountNoun 0 Finger)

• Mt AnimalPhysiologyMt
• -(conceptuallyRelated Fingernail Finger)
  (properPhysicalPartTypes Hand Finger)
  (relationAllInstance age Finger       
  (YearsDuration 0 200)) (relationAllInstance
  widthOfObject Finger        (Meter 0.001 0.2))
  (relationAllInstance heightOfObject Finger
         (Meter 0.001 0.2)) (relationAllInstance
  lengthOfObject Finger        (Meter 0.01 0.5))
  (relationAllInstance massOfObject Finger       
  (Kilogram 0.001 1))

7
Bootstrapping with ResearchCyc

• Cyc has vocabulary about objects in the world and
  relationships
• Cyc could still use more common relationships
• BURC uses what Cyc already has lots of parsed
  text to create new Cyc entries for common
  relationships found in the text
• Lenats Bootstrap Hypothesis once Cyc reaches a
  certain level/scale it can help in its own
  development and start using NLP to augment its
  knowledge base
• BURC should help test this hypothesis

8
The BURC Process From seedsHypothe-seeds

• Use the link grammar parser for bulk parsing of
  text, primarily narratives based in worlds like
  ours. Other text styles could be included.
• Operates in two directions
• Forward from text to CycL
• Backwards from existing CycL to the text to find
  new forward patterns

9
BURC Process - 2

• Load the link fragments into a database (1 and 2
  link fragments), and compute frequency of
  fragment occurrences. The database will be in a
  SQL format so multiple queries can be formed
  dynamically.
• Using Cyc knowledge as a starting point (the
  seeds), extract knowledge for use in Cyc
• Given a set of seed facts in Cyc, identify how
  those facts are represented as link fragments in
  the database
• Generate conjectures as to new knowledge AND new
  knowledge extraction patterns using the fragment
  patterns.

10
BURC Process - 3

• Use Cyc knowledge directly to conjecture new
  statements
• Cyc has lexical knowledge, which can be used as
  templates against the DB to form new statements
• For example, common adjectives applied to noun
  classes
• Cyc knows WhiteColor and Blouse but does not
  know that white is a common blouse color,
  although it becomes apparent after reading some
  text
• Optionally, gather supporting background
  statistics for hypothesis verification using
  other sources
• Perhaps Google desktop with a larger than fully
  parsed corpus
• Perhaps check against answer extraction engines

11
KNEXT (KNowledge EXtraction from Text)

• Deriving general world knowledge from texts and
  taxonomies
• http//www.cs.rochester.edu/schubert/projects/wor
  ld-knowledge-mining.html
• Lenhart K. Schubert and Matthew Tong, "Extracting
  and evaluating general world knowledge from the
  Brown Corpus", Proc. of the HLT-NAACL Workshop on
  Text Meaning, May 31, 2003, Edmonton, Alberta,
  pp. 7-13.
• System extracts commonsense relationships from
  text
• Limited to the pre-parsed Penn Treebank
• Generated 117,326 propositions (about 2 per
  sentence)
• About 60 judged reasonable by any given judge

12
KNEXT (Example)

• (BLANCHE KNEW 0 SOMETHING MUST BE CAUSING STANLEY
  'S NEW, STRANGE BEHAVIOR BUT SHE NEVER ONCE
  CONNECTED IT WITH KITTI WALKER.)
• A FEMALE-INDIVIDUAL MAY KNOW A PROPOSITION.
• SOMETHING MAY CAUSE A BEHAVIOR.
• A MALE-INDIVIDUAL MAY HAVE A BEHAVIOR.
• A BEHAVIOR CAN BE NEW.
• A BEHAVIOR CAN BE STRANGE.
• A FEMALE-INDIVIDUAL MAY CONNECT A
  THING-REFERRED-TO WITH A FEMALE-INDIVIDUAL.
• ((I (Q DET FEMALE-INDIVIDUAL) KNOWV (Q DET
  PROPOS))
• (I (F K SOMETHINGN) CAUSEV (Q THE
  BEHAVIORN))
• (I (Q DET MALE-INDIVIDUAL) HAVEV (Q DET
  BEHAVIORN))
• (I (Q DET BEHAVIORN) NEWA)
• (I (Q DET BEHAVIORN) STRANGEA)
• (I (Q DET FEMALE-INDIVIDUAL) CONNECTV (Q
  DET THING-REFERRED-TO)
• (P WITHP (Q DET FEMALE-INDIVIDUAL))))

13
Other Extraction Pattern Research

• Towards Terascale Knowledge Acquisition (Pantel,
  Ravichandran and Hovy, 2004)
• Learning Surface Text Patterns for a Question
  Answering System (Ravichandran Hovy, 2002)
• Defined Pattern Precision P Ca/Co
• Ca total number of patterns with answer term
  present
• Co Total number of patterns with any term
  present
• DIRT Discovery of Inference Rules from Text
  (Lin Pantel, 2001)

14
Other Lexical Knowledge Research

• VerbOcean (Chklovski Pantel) Collecting pairs
  and searching to verify relationships
• Lexical Acquisition via Constraint Solving
  (Pedersen Chen) Acquiring syntactic and
  semantic classification rules of unknown words
  for LGP
• Information Extraction Using Link Grammar papers
• Automatic Meaning Discovery Using Google

15
The General Backwards Model

• Given some Cyc relation Pred(?X,?Y)
• Create SQL search query
• Lookup in Cyc lexical entries for X Y ? LX, LY
• Select from LGPTable where Term1"ltLXgt" and
  Term3"ltLYgt
• System returns records LX Link1 Term2
  Link2 LY (Freq)
• Generate new hypothetical extraction patterns
• Select from LGPTable where Link1"ltL1gt" and
  Link2"ltL2gt" and Term2"ltT2gt
• L1 T2 L2 ? generate hypothetical record (
  Pred ?S1?S3 )
• Frequency information is propagated forward

16
The General Backwards Model - 2

• Optional Search Cyc for ?PRED (X,Y) and use the
  set to form a local ambiguity class to reduce
  search labor and identify ambiguity. One rule ?
  multiple relations.
• Stored as SQLTemplate \ Pattern \
  Pred1/Pred2//PRedN
• Need to explore (canidateBinaryPred ARG1 ARG2
  RELN)
• Optional Form more specific patterns for
  Pred(X,_) and Pred(_,Y)

17
Update the LGParsers CycL Rules

• ltrulegt
• ltpatterngt Link1 Term2 Link2lt/patterngt
• ltdefinegt?ITEMr lt/definegt
• ltbodygt(is-node ?ITEMr "R")lt/bodygt
• ltdefinegt?ITEMl lt/definegt
• ltbodygt(is-node ?ITEMl "R")lt/bodygt
• ltbodygt(?PRED1 ?ITEMl ?TERMr)lt/bodygt
• ...
• ltbodygt(?PREDN ?ITEMl ?TERMr)lt/bodygt
• lt/rulegt

• There are rules for translation of LGP output
  into CycL
• If the frequency information warrants it then we
  can generate new LGP rules
• Results in expanded parser precision

18
Forward Mining Adjective Relations

• There are 1941 GAFs on adjSemTrans, the primary
  lexical adjective predicate
• Find applicable fragments and use definitions
• Select from LGPTable Where NumLinks1 and
  Link1'a' and Term1 like '.a' and Term2 like
  '.n
• Returns records Term1.a a Term2.n
• Potentially test using either an internal or
  search engine based relevancy metric
• Query Cyc for (adjSemTrans ltterm1gt-TheWord ?N
  RegularAdjFrame (?Pred NOUN ?Val))
• Generate (plausiblePredValOFType ltterm2gt lt?Predgt
  lt?Valgt)
• Possibly generate parsing rule

19
Mining Adjective Knowledge Example

• white blouse as factoid
• white.a a blouse.n
• Potentially test using an internal or search
  engine relevancy metric GC70400
• (adjSemTrans White-TheWord 11 RegularAdjFrame
  (mainColorOfObject NOUN WhiteColor))
• Hypothesis (plausiblePredValueOfType Blouse
  mainColorOfObject WhiteColor)

20
Update the LGParsers CycL Rules - 2

• ltrulegt
• ltpatterngt Term1.a a lt/patterngt
• ltdefinegt?ITEMr lt/definegt
• ltbodygt(is-node ?ITEMr "R")lt/bodygt
• ltdefinegt?ITEMl lt/definegt
• ltbodygt(?PRED ?ITEMr ?VAL)lt/bodygt
• lt/rulegt

• There are rules for translation of LGP output
  into CycL
• We can use the adjSemTrans data to generate new
  translation rules
• Results in expanded parser precision

• ltrulegt
• ltpatterngt white.a a lt/patterngt
• ltdefinegt?ITEMr lt/definegt
• ltbodygt(is-node ?ITEMr "R")lt/bodygt
• ltdefinegt?ITEMl lt/definegt
• ltbodygt(mainColorOfObject ?ITEMr
  WhiteColor)lt/bodygt
• lt/rulegt

21
Mined Finger Descriptions

• 000010(plausiblePredValueOfType Finger
  feelsSensation (PositiveAmountFn
  LevelOfSoreness))
• 000037(plausiblePredValueOfType Finger
  forceCapacity Strong)
• 000025(plausiblePredValueOfType Finger
  forceCapacity Strong)
• 000025(plausiblePredValueOfType Finger
  hardnessOfObject Hard)
• 000037(plausiblePredValueOfType Finger
  hardnessOfObject (MediumToVeryHighAmountFn
  Hardness))
• 000037(plausiblePredValueOfType Finger
  hardnessOfObject (MediumToVeryHighAmountFn
  Hardness))
• 000002(plausiblePredValueOfType Finger
  hasEvaluativeQuantity (MediumToVeryHighAmountF
  n Goodness-Generic))
• 000002(plausiblePredValueOfType Finger
  hasPhysicalAttractiveness GoodLooking)
• 000047(plausiblePredValueOfType Finger isa
  (LeftObjectOfPairFn REPLACE))
• 000015(plausiblePredValueOfType Finger isa
  (RightObjectOfPairFn REPLACE))
• 000155(plausiblePredValueOfType Finger
  lengthOfObject (RelativeGenericValueFn
  lengthOfObject REPLACE highAmountOf))
• 000155(plausiblePredValueOfType Finger
  lengthOfObject (RelativeGenericValueFn
  lengthOfObject REPLACE highToVeryHighAmountOf
  ))
• 000003(plausiblePredValueOfType Finger
  mainColorOfObject BlackColor)
• 000010(plausiblePredValueOfType Finger
  mainColorOfObject LightYellowishBrown-Color)
• 000010(plausiblePredValueOfType Finger
  mainColorOfObject ModerateYellowishBrown-Color
  )
• 000010(plausiblePredValueOfType Finger
  mainColorOfObject SunTan-FleshColor)
• 000002(plausiblePredValueOfType Finger
  possessiveRelation SuddenChange)

22
Mined Finger Descriptions

• 000006(plausiblePredValueOfType Finger
  possessiveRelation (HighAmountFn Speed))
• 000094(plausiblePredValueOfType Finger
  rigidityOfObject (HighAmountFn Rigidity))
• 000060(plausiblePredValueOfType Finger
  sizeParameterOfObject (RelativeGenericValueFn
  sizeParameterOfObject REPLACE highAmountOf))
  
• 000052(plausiblePredValueOfType Finger
  sizeParameterOfObject (RelativeGenericValueFn
  sizeParameterOfObject REPLACE
  highToVeryHighAmountOf))
• 000060(plausiblePredValueOfType Finger
  sizeParameterOfObject (RelativeGenericValueFn
  sizeParameterOfObject REPLACE
  highToVeryHighAmountOf))
• 000285(plausiblePredValueOfType Finger
  sizeParameterOfObject (RelativeGenericValueFn
  sizeParameterOfObject REPLACE
  veryLowToLowAmountOf))
• 000074(plausiblePredValueOfType Finger
  sizeParameterOfObject (RelativeGenericValueFn
  sizeParameterOfObject REPLACE
  veryLowToLowAmountOf))
• 000029(plausiblePredValueOfType Finger
  speedOfObject-Underspecified (LowAmountFn
  Speed))
• 000138(plausiblePredValueOfType Finger
  surfaceFeatureOfObj Slippery)
• 000074(plausiblePredValueOfType Finger
  temperatureOfObject Warm)
• 000004(plausiblePredValueOfType Finger
  textureOfObject Rough)
• 000168(plausiblePredValueOfType Finger
  thicknessOfObject (RelativeGenericValueFn
  thicknessOfObject REPLACE highAmountOf))
• 000168(plausiblePredValueOfType Finger
  thicknessOfObject (RelativeGenericValueFn
  thicknessOfObject REPLACE highToVeryHighAmoun
  tOf))
• 000182(plausiblePredValueOfType Finger
  wetnessOfObject Wet)

23
Verb Semantic Filtering -1Discovering what a
finger can do

• A similar process can be used finding information
  based on verb semantic parsing frames
• For each potential ltNOUNWORDgt-ltVERBgt pair query
  Cyc to find basic relationships using the verb
  semantic templates
• (and
• (denotation ltNOUNWORDgt ?NOUNTYPE ?N ?CYCTERM)
• (wordForms ?WORD ?PRED ""ltVERBgt"")
• (speechPartPreds ?POS ?PRED)
• (semTransPredForPOS ?POS ?SEMTRANSPRED)
• (?SEMTRANSPRED ?WORD ?NUM ?FRAME ?TEMPLATE))
• Verify for each potential relationship (ltSPREDgt
  ltVERTERMgt ltCYCTERMgt) derivable from ?TEMPLATE
  that it makes sense in the ontology
• (and
• (arg1Isa ltSPREDgt ?VTYP)
• (arg2Isa ltSPREDgt ?CTYP)
• (genls ltCYCTERMgt ?CTYP)
• (genls ltVERBTERMgt ?VTYP) )

24
Verb Semantic Filtering -2Templates of Movement

• (verbSemTrans Move-TheWord 0 IntransitiveVerbFram
  e        (and            (isa ACTION
  MovementEvent)            (primaryObjectMoving
  ACTION SUBJECT)))
• (verbSemTrans Move-TheWord 1 IntransitiveVerbFram
  e        (and            (isa ACTION
  ChangeOfResidence)            (performedBy
  ACTION SUBJECT)))
• (verbSemTrans Move-TheWord 2 TransitiveNPFrame
         (and            (isa ACTION
  CausingAnotherObjectsTranslationalMotion)
             (objectActedOn ACTION OBJECT)
             (doneBy ACTION SUBJECT)))
• (arg1Isa performedBy Action)
• (arg2Isa performedBy Agent-Generic)

25
Verb Semantic Filtering - 3

• BURC can use Cycs knowledge of what things can
  perform what actions or have what attributes to
  filter out implausible relationships.
• (behaviorCapableOf Finger CausingAnotherObje
  ctsTranslationalMotion doneBy)
• (behaviorCapableOf Finger ChangeOfResidence
  performedBy)
• (behaviorCapableOf Finger Inspecting
  performedBy)
• (behaviorCapableOf Finger Movement-Translati
  onEvent primaryObjectMoving)
• (behaviorCapableOf Finger MovementEvent
  primaryObjectMoving)
• (behaviorCapableOf Finger PushingAnObject
  providerOfMotiveForce)
• (behaviorCapableOf Finger Sliding-Generic
  objectMoving)
• (behaviorCapableOf Finger Sliding-Generic
  primaryObjectMoving)
• (behaviorCapableOf Finger Slipping
  objectMoving)
• (behaviorCapableOf Finger Slipping
  primaryObjectMoving)
• Cyc can help in its own knowledge entry process.
  62 of generated hypothesis were filtered out
  using semantic role filtering.

26
Other Direct Extraction Rules

• Some underspecified patterns exist just based
  on the links
• This could be used to extract ConceptNet like
  output directly from link records
• Examples
• ltobj1gtssltactgt.vosltobj2gt ? capableOf(ltobj1gt,
  ltactgt ltobj2gt)
• ltactgt.v osltobjgt ? CapableOfReveivingAction(ltob
  jgt,ltactgt)
• ltobjgtsltactgt.v ? capableOf(ltobjgt,ltactgt)

27
Quest for Metrics

• Percentage of hypothesis that make sense to a
  panel of judges
• Percentages of hypothesis that are already known
  to Cyc
• Percentage of hypothesis that are known in other
  knowledge sources (WordNet, Sumo/Milo, VerbOcean,
  MIT OpenMind)
• Number of hypothesis generated vs. number of
  records
• What percentage of relations in Cyc can be found
  in the fragment pool
• The Pattern Precision measure
• Maybe compare against KNEXT but need to see if
  they return real numbers
• Unfortunately we dont know all possible
  knowledge (otherwise we wouldnt be doing this),
  because if we did we could measure recall and
  precision.
• Simple space estimate (2.3K binary predicates
  85K constants 85K constants 16.617500 T
  simple possibilities)

28
Desired Outputs

• Version of link grammar for bulk reading and
  generating fragments
• Database control program to queue texts, monitor
  their processing, and merge the fragment results
• The database of fragments with fragment counts
  for some corpus
• The hypothesis set generated by the system
• Optionally an OpenMind / ConceptNet like set of
  commonsense factoids
• Open enough that others could duplicate

29
Did any of that make sense?

• Comments?
• Questions?
• Suggestions?