30 Y ears of Higher-Order Unification
G ´erard Huet
INRIA
TPHOL2002,Hampton
-1-
35 Y ears of Higher-Order Unification
G ´erard Huet
TPPHOL2007,Somefarawayplace
-2-
40 Y ears of Higher-Order Unification
Not me!
TPPHOLT2012,TroulesOies
-3-
T o o m uc h is to o m uc h
\begin{gripe}
•Toomanypapersontoonarrowtopics
•Notenoughthematicmobility
•Notenoughinterdisciplinarywork
•ToomuchBrownianMotionResearch
\end{gripe}
-4-
Mating
\begin{benediction}
•AtLICSinSantaCruz,DanaScottpreachedtheWeddingofLogicandCS
•TodayIpreachFreeLoveamongDisciplines
\end{benediction}
-5-
Preac hing (in the desert)
Warning.Adviceintendedforstudentsandyoungscientists.
\begin{preach}
•Changeresearchtopicevery5years
•Changedisciplineatleastonceinyourcareer
•Open-mindedness
•SeekApplicationsfortheJoyofit
•Dowhatonlyyoucando
•NesoyezpasunmoutondePanurge
•Don’ttrytoownatopic
\end{preach}
-6-
Bet w een Logic and Algebra
•LinearLogicasLogicalFoundations
•NoContraction,noWeakening
•SequentsasBagsofLiterals
•LambekSyntacticCalculus\/
•NonCommutativeLL
•NonAssociativeLL
•Pregroups
-7-
What I attempted to do
•Acommentedbibliographyonthetopic
•Withsomereferencetoapplications
•Whathappened
-8-
In a n utshell
•Dowek.Handbookofautomatedtheoremproving.
•Miller’spatternsandmixedprefixes
•λProlog
•DependentTypes
•Decidabilityofrestrictedmatching
•Mixingβandequations
•RewriteLogic,ExplicitSubstitutions
•OogedeMoor&Sittampalam.Matchingorder2.5.
•Loader.Higher-orderβmatchingisundecidable.
-9-
The main results in m y PhD
•Undecidability
•Non-wellfoundedness
•Constraints
•Pre-unifiers
-10-
One of m y T ren to slides
Non-deterministicprogrammingisnobigdeal.WhyshouldyousurrendercontroltoaPROLOGblackbox?
Thethreegoldenrulesofnon-deterministicprogramming:
•Identifywellyoursearchstatespace
•Representstatesasnon-mutabledata
•Provetermination
Thelastpointisessentialforunderstandingthegranularityandenforcingcompleteness.
-11-
The main results in m y Do ctorat d’Etat
•CorrectdefinitionofCSU
•CSMPre-unifiers
•Second-ordermatching
•Implementation
•PrincipalUnifiersofRegularTerms
-12-
The main op en problems at the time
•Second-orderunification
•Associativeunification
•3rdandhigherordermatching
-13-
Where to find HO terms
•inmathematicaldescriptions
•inlogicalproofs
•incomputerprograms
•insemanticdescriptionsofnaturallanguage
•inDNApatterns?
•in...
-14-
Applications
•Proofassistantsandtheoremprovers
•Softwareengineering
•Computationallinguistics
•Health?
-15-
Computational Linguistics and HOU
•MontagueSemantics
•Anaphoraresolution
•Pronouns
•Focus
•Ellipsis
•Accent
•Colors
-16-
Anaphora Resolution
•JohnlikesMaryandPeterdoestoo.
•like(j,m)∧R(p)
•R(j)=like(j,m)
•R←λx.like(x,m)
-17-
More complex examples
•JohnlikeshiswifeandPetertoo.
•JohnsaysthathelikeshismotherandPeterdoestoo.
•JohnspenthispaycheckbutPetersavedit.
•MarywantstogotoSpainandFredwantstogotoPerubutbecauseoflimitedresources,onlyoneofthemwill.
-18-
Sp eec h and Linguistics
•TheWatergatetapes
•Languageasaplaypenforlinguists
•TypicalpapersinCL
•Non-wellfoundednessofpublications
-19-
Promising future in vestigations ?
•Proofnets&GoI
•ComputationalMathematics
•Constraintsengines
•Cayenne
•LiterateProgramming
•Understandingnaturallanguage
•Understandinggenes
•CognitiveScience
-20-