ComputerScience:TheEver-ExpandingSphere
EdLazowskaBill&MelindaGatesChairin
ComputerScience&EngineeringFoundingDirector,eScienceInstitute
UniversityofWashington
Dean’sSeminarSeries,McCormickSchoolofEngineeringNorthwesternUniversity
April2016
http://lazowska.cs.washington.edu/NU.pdf,pptx
Today
• AreminderoftheextraordinaryprogressthatComputerSciencehasachieved
• Aglimpseatwhat’shappeningtoday• A21st centuryviewofthefield• TheroleofComputerScienceinthemodernuniversity
– Onequickexample:TheUniversityofWashington eScienceInstitute• TheroleofComputerScienceinthemodernworld• Studentresponse(enrollmenttrends)• Institutionalresponse
1969– Fortysevenyearsago…
Credit: PeterLee,MicrosoftResearch
Credit: PeterLee,MicrosoftResearch
Credit: PeterLee,MicrosoftResearch
Credit: PeterLee,MicrosoftResearch
Credit: PeterLee,MicrosoftResearch
Withnearly5decadesofhindsight,whichhadthegreatestimpact?
• Unlessyou’rebigintoTang* andVelcro* (orsexanddrugs),theanswerisclear…
• Andsoisthereason…
EXPONENTIALS US*Commonly– althougherroneously – attributed tothespaceprogram
• Constantcapabilityatexponentiallydecreasingcost• Exponentiallyincreasingcapabilityatconstantcost
Theexponentialimprovementsthathavecharacterizedcomputingcanbeexploitedintwoways
StoragePrice/MB,USD(semi-logplot)
MicroprocessorPerformance,MIPS(semi-logplot)
JohnMcCallum /Havard Blok
Disk
RAM
Flash
RayKurzweil
1955196019651970197519801985199019952000200520102015 1970197519801985199019952000 2005
Today,theseexponentialimprovementsintechnology(andalsoinalgorithms!)areenablinga“bigdata”revolution
• Aproliferationofsensors– Thinkaboutthesensorsonyourphone
• Moregenerally,thecreationofalmostallinformationindigitalform– Itdoesn’t needtobetranscribedinordertobeprocessed
• Dramaticcostreductionsinstorage– Youcanaffordtokeepallthedata
• Dramaticincreasesinnetworkbandwidth– Youcanmovethedatatowhereit’sneeded
• Dramaticcostreductionsandscalabilityimprovementsincomputation– WithAmazonWebServices,1000computersfor1daycoststhesameas1computerfor
1000days
• Dramaticalgorithmicbreakthroughs– Machinelearning,datamining– fundamentaladvancesincomputerscienceand
statistics
• Evermorepowerfulmodelsproducingever-increasingvolumesofdatathatmustbeanalyzed
So,exactlywhat’smeantby“bigdata”?
Credit: DanAriely,DukeUniversity
Seriousanswer:“bigdata”isenablingcomputerscientiststoputthe“smarts” intoeverything
• Smarthomes• Smartcars• Smarthealth• Smartrobots• Smartcrowdsandhuman-computersystems• Smarteducation• Smartinteraction(virtualandaugmentedreality)• Smartcities• Smartdiscovery
Business + Technology in the Exponential Economy
Smarthomes(theleafnodesofthesmartgrid)
ShwetakPatel,UniversityofWashington2011MacArthurFellow
Smartcars
DARPAGrandChallenge DARPAUrbanChallenge GoogleSelf-DrivingCar
Adaptivecruisecontrol Self-parkingTeslaModelS
P4medicine
Smarthealth
Evidence-basedmedicineLarrySmarr – “quantifiedself ”
Smartrobots
Smartcrowdsandhuman-computersystems
ZoranPopovicUWComputerScience&Engineering
DavidBakerUWBiochemistry
Smarteducation
ZoranPopovicUWComputerScience&Engineering
Smartinteraction
Smartinteraction– contentcreation
SteveSeitzUWComputerScience&Engineering+GoogleSeattle
Smartcities
Smartdiscovery:“TheFourthParadigm”
1. Empirical+experimental2. Theoretical3. Computational4. Data-Intensive
JimGray
Eachaugments,vs.supplants, itspredecessors– “anotherarrowinthequiver”
Energy&Sustainability
Security,Privacy,&Safety
Advancing theDevelopingWorld
Medicine&GlobalHealth
Education
ScientificDiscovery
Transportation
NeuralEngineering
ElderCareAccessibility
Interactingwith thePhysicalWorld:“TheInternet ofThings”
mobilecomputing
robotics
computervision
machinelearning
humancomputerinteraction
datascience
sensors
naturallanguageprocessing CORECSE
AI,systems,theory,languages,
etc.
cloudcomputing
TechnologyPolicyandSocietalImplications
A21st centuryviewofComputerScience:Afielduniqueinitssocietalimpact
Energy&Sustainability
Security,Privacy,&Safety
Advancing theDevelopingWorld
Medicine&GlobalHealth
Education
ScientificDiscovery
Transportation
NeuralEngineering
ElderCareAccessibility
Interactingwith thePhysicalWorld:“TheInternet ofThings”
TechnologyPolicyandSocietalImplications
Isthisstuffcomputerscience?
“Thelastelectricalengineer”
“Iamworriedaboutthefutureofourprofession.…Iseetheworldasaninvertedpyramid.Itbalancesprecariouslyonthenarrowpointatthebottom.…Thispoint isbeing impressedintotheground bytheheavyweightatthewidetopoftheinvertedpyramidwherealltheapplicationsreside.…Electricalengineering willbeindangerof shrinking intoaneutronstarofinfiniteweightandimportance,butinvisibletotheknownuniverse.…SomewhereinthebasementofInteloritssuccessor…thelastelectricalengineerwillsit.”
BobLuckyIEEESpectrumMay1998
Credit: AlfredSpector, Google(ret.)
“ComputerScience:Theever-expandingsphere”
Energy&Sustainability
Security,Privacy,&Safety
Advancing theDevelopingWorld
Medicine&GlobalHealth
Education
ScientificDiscovery
Transportation
NeuralEngineering
ElderCareAccessibility
Interactingwith thePhysicalWorld:“TheInternet ofThings”
mobilecomputing
robotics
computervision
machinelearning
humancomputerinteraction
datascience
sensors
naturallanguageprocessing CORECSE
AI,systems,theory,languages,
etc.
cloudcomputing
TechnologyPolicyandSocietalImplications
“ComputerScience:Theever-expandingsphere”
AtUW,we’vebeeninvestinginallareas:• We’veaugmented
thecore• Wehaveaserious
storyinallareasofthe“connections”ringandnearlyallofthesocietalchallengeareas
TheroleofComputerScienceinthemodernuniversity
• TheCenterforSensorimotorNeuralEngineering, anNSFEngineering ResearchCenter
• TheCenterforGameScience,fundedbytheGatesFoundationandDARPAtocreaterevolutionarygamesforscientificdiscoveryandforlearning
• TheeScienceInstitute,funded bytheMoore,Sloan,WashingtonResearch,andNationalScienceFoundations tobringadvancesindata-intensivediscoverytoresearcherscampus-wide
• dub – “design-use-build”– acampus-widecollaborationthathasmadeUWoneof thetopinstitutions inthenationinhuman-computer interaction
• Urban@UW,acampus-wideurbansciencecollaboration
• TheTaskarCenterforAccessibleTechnologydevelopsanddeploys technologies thatincreaseindependence andimprovequalityoflifeforindividualswithmotorandspeechimpairments
• Change,acampus-widecollaborationexploringhowtechnologycanimprovethelivesofunderserved populations inlow-incomeregions
• TheTechPolicyLab,ajointeffortofCSE,theSchoolof Law,andtheInformationSchool, fundedbyMicrosoft
• GIX – theGlobalInnovationExchange– anewkindofeducationthatisglobal,project-based,andintegratestechnology, design,andentrepreneurship
• TheIntelScienceandTechnologyCenterforPervasiveComputing,ledbyUW,withresearchersfromCornell,GeorgiaTech,Rochester,Stanford,andUCLA
“Allacrossourcampus,theprocessofdiscoverywillincreasinglyrelyonresearchers’abilitytoextractknowledgefromvastamountsofdata...Inordertoremainattheforefront,UWmustbealeaderinadvancingthesetechniquesandtechnologies,andinmaking[them]accessibletoresearchersinthebroadestimaginablerangeoffields.”[2007conceptpaper]
Onequickexample:TheUniversityofWashingtoneScienceInstitute
• UniversityofWashington– $725,000/yearforstaffsupport– $600,000/yearforfacultysupport
• NationalScienceFoundation– $2.8millionover5yearsforgraduateprogramdevelopmentand
Ph.D.studentfunding(IGERT)
• GordonandBettyMooreFoundationandAlfredP.SloanFoundation
– $37.8millionover5yearstoUW,Berkeley,NYU
• WashingtonResearchFoundation– $9.3millionover5yearsforfacultyrecruitingpackages,postdocs
• Also$7.1milliontotheclosely-alignedInstituteforNeuroengineering
Majorsourcesofsupportforour“coreeffort”
• TheFoundationshaveafocusonnoveladvancesinthephysical,life,environmental,andsocialsciences
• Theyrecognizedtheemergenceofdata-intensivediscoveryasanimportantnewapproachthatwouldleadtonewadvances
• Theyperceivedanumberofimpedimentstosuccess• Theysoughtpartnerswhowerepreparedtoworktogetherina
distributedcollaborative experiment focusedontacklingtheseimpediments
GenesisoftheMoore/Sloan“DataScienceEnvironments”effort
Vision
EdLazowskaCSE
Datasciencemethodology
Lifesciences Environmentalsciences
Socialsciences Physicalsciences
CeciliaAragonHumanCenteredDesign&Engr.
MagdaBalazinskaComputerScience&Engineering
CarlosGuestrinCSE
BillHoweCSE
RandyLeVequeAppliedMathematics
WernerStuetzleStatistics
TomDanielBiology
GingerArmbrustOceanography
AndyConnollyAstronomy
JohnVidaleEarth&SpaceSciences
JoshBlumenstockiSchool
MarkEllisGeography
TylerMcCormickSociology,Statistics,CSSS
ThomasRichardsonStatistics,CSSS
EmilyFoxStatistics
JeffHeerCSE
BillNobleGenomeSciences
DavidBeckChemicalEngr.
Ouroriginalcorefacultyteam(muchexpandednow)
EdLazowskaCSE
Datasciencemethodology
Lifesciences Environmentalsciences
Socialsciences Physicalsciences
CeciliaAragonHumanCenteredDesign&Engr.
MagdaBalazinskaComputerScience&Engineering
CarlosGuestrinCSE
BillHoweCSE
RandyLeVequeAppliedMathematics
WernerStuetzleStatistics
TomDanielBiology
GingerArmbrustOceanography
AndyConnollyAstronomy
JohnVidaleEarth&SpaceSciences
JoshBlumenstockiSchool
MarkEllisGeography
TylerMcCormickSociology,Statistics,CSSS
ThomasRichardsonStatistics,CSSS
EmilyFoxStatistics
JeffHeerCSE
BillNobleGenomeSciences
DavidBeckChemicalEngr.
Ouroriginalcorefacultyteam(muchexpandednow)
• Educationalinitiatives attheBachelors,Masters,andDoctorallevels,plusprofessionaleducationcertificateprogramsandCourseraMOOCs
• Hiresof“pi-shaped”faculty(facilitatedbytheProvostandtheWashingtonResearchFoundation)
Amongouractivities
• Avibrantprogramofdual-mentoredpostdocsandgraduatestudents
• ApermanentstaffofsuperbPh.D.-levelDataScientistsspanningdisciplines
• Acollaboratory:theWRFDataScienceStudio• Myriadtrainingandmentoringactivities:
short-courses,workshops,officehours• Deeppartnerships,plusan“incubation
program”fordata-intensiveresearchprojects
Microbialcommunity visualizedwithDNAstain
100μm
Challenges:• Integrationacrossdifferentdatatypes• Distributedandremotelabs
Oneexampleofadeeppartnership
Roleofmicrobesinmarineecosystems• GingerArmbrust(Oceanography), BillHowe(CSE+eScienceInstitute)
Credit: GingerArmbrust, UniversityofWashington
Credit: GingerArmbrust, UniversityofWashington
Queryacrossdatasets inreal-time:“notjustfaster…different!”
DanHalperin,ResearchScientist,eScienceInstitute
KonstantinWeitzGraduatestudent,CSE
Credit: GingerArmbrust, UniversityofWashington
Integrating acrossphysics,biology,andchemistry
SatellitelinkSeaFlow instrument Labcomputer
Ship computer
Processeddata
LabcomputerCloud – SQLShare
Webdisplay –collaboratorcomputers
Othershipdatastreams
automated
manual
Completelyautomated
Credit: GingerArmbrust, UniversityofWashington
Connecting acrossdistributed labs
• Thereareover4,000homeless familiesintheTri-countyareaeveryyear
• Familiesspendonaverage8monthsmovingfromsheltertoshelter
• GoaloftheBill&MelindaGatesFoundation andBuildingChanges:Cutfamilyhomelessness inhalfby2020andreducethetimeafamilyspends homeless to1month
InSnohomish, Pierce,andKingcounties
PredictorsofPermanentHousing forHomelessFamilies• Projectleads:NeilRocheandAnjanaSundaram, TheBillandMelinda
GatesFoundation• DSSGFellows: Fablina Ibnat,JasonPortenoy, Chris Suberlak, JoanWang• ALVAstudents:CameronHolt,Xilalit Sanchez• DataScientistMentors:ArielRokem,Bryna Hazelton
Oneofdozensofincubationprojects
Credit: Fablina Ibnat,UniversityofWashington
Homelessfamiliesmaytakemanypathwaysthroughprograms
Emergency shelter
Transitional housing
Rapid re-housing
Permanent housing
Housing with services Unsuccessful
exitCredit: Fablina Ibnat,UniversityofWashington
KingHMISextract
PierceHMISextract
SnohomishHMISextract
Cleaneddataforhouseholds
Cleaneddataforfamilies
Familyenrollments
Mappingofenrollmentstoepisodes
Familyepisodes
Createdatacleaningpipeline formessyHomelessManagement InformationSystemsdata
Credit: Fablina Ibnat,UniversityofWashington
Developvisualizationstoshowhowhomelessfamiliesmovethrough programs
Credit: Fablina Ibnat,UniversityofWashington
Conduct analysistoidentifypredictorsofpermanenthousing
Correlation with successful outcome, by family characteristics
Correlation with successful outcome, by homelessness program
Emergency Shelter use tends to be associated with unsuccessful outcomes (unsurprising!)
Homelessness Prevention programs more strongly associated with positive outcomes than transitional housing
Substance abuse strongly associated with unsuccessful outcomes
Parent employment strongest predictor of successful outcomes
Credit: Fablina Ibnat,UniversityofWashington
TheroleofComputerScienceinthemodernworld
1. Every21st centurycitizenneedstohavefacilitywith“computationalthinking”– problemanalysisanddecomposition(stepwiserefinement),abstraction,algorithmicthinking,algorithmicexpression,stepwisefaultisolation(debugging),modeling– Computational thinking isnot“thisparticular
operatingsystem”or“thatparticularprogramminglanguage”
– Computational thinking isnotevenprogramming.It’samodeofthought– awayofapproaching theworld
– Programming isthehands-on, inquiry-basedwaythatweteachcomputationalthinking andtheprinciplesofcomputer science
2. FieldsfromAnthropologytoZoologyarebecominginformation fields,andthatthosewhocanbendthepowerofthecomputertotheirwill–computationalthinking,butalsocomputerscienceingreaterdepth– willbepositionedforgreatersuccessthanthosewhocan’t– Datascienceisaperfectexample
3. PrettymuchalloftheSTEMjobsareincomputerscience– Inthecomputing industry, whichisnotDilbert– it’san
intellectuallyexciting,highlycreativeandinteractive,“powertochangetheworld”field
– Inallsortsofother fieldswherepeopleeducatedascomputer scientists– notmerelypeoplewithsomecomputer sciencebackground– areessential
73%
10%
3%3%
5%
6%
JobGrowth,2014-24- U.S.BureauofLaborStatistics
Computer occupations (15-1100)
Engineers(17-2000)
Lifescientists (19-1000)
Physical scientists (19-2000)
Socialscientists andrelatedworkers(19-3000)
Mathematical science occupations (15-2000)
STEMjobgrowth,2014-24(USBureauofLaborStatistics)
Data from the spreadsheet at http://www.bls.gov/emp/ind-occ-matrix/occupation.xlsx
55%
26%
6%
4%
5%4%
JobOpenings(Growth+Replacement),2014-24- U.S.BureauofLaborStatistics
Computer occupations (15-1100)
Engineers(17-2000)
Lifescientists (19-1000)
Physical scientists (19-2000)
Socialscientists andrelatedworkers(19-3000)
Mathematical science occupations (15-2000)
Datafromthespreadsheetathttp://www.bls.gov/emp/ind-occ-matrix/occupation.xlsx
STEMjobopenings(growth+replacement),2014-24(USBLS)
Data from the spreadsheet at http://www.bls.gov/emp/ind-occ-matrix/occupation.xlsx
16%ofallnewwages,acrossall fields
Credit: Code.org
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
CurrentAnnualCompletions
AdditionalAnnualCompletionsNeeded,2016-21
ComputerScience
Engineering
HealthProfessions*
Research,Science,
Technical*
*Gapexistsatthegraduateand/orprofessionallevelonly
HighDemandFieldsinWashingtonState,BaccalaureateLevel&AboveWSAC/SBCTC/WTECB,October2013
Data from Table 2 at http://www.wsac.wa.gov/sites/default/files/2013.11.16.Skills.Report.pdf
FieldswithworkforcegapsinWashingtonState(Baccalaureatelevelandabove)
KingCountyWA’sAerospaceWorkforceQ: What field has the largest total
number of current employees in King County’s aerospace industry?
Q: What field has the greatest predicted number of new employees needed by King County’s aerospace industry from 2013-2023?
Q: What field has the greatest predicted compound annual growth rate for King County’s aerospace industry from 2013-2023?
Q: What field has the greatest predicted annual gap between supply and demand for King County’s aerospace industry from 2013-2023 (where “supply” is not “degrees granted” but rather the industry’s current ability to hire)?
A: Computer Science
Studentsarerespondingtoallthreeimperatives
1. Demandforintroductorycoursesisbooming
2. Demandforupper-divisionandgraduatecoursesbynon-majorsisbooming
3. Demandforthemajorisbooming
0
500
1000
1500
2000
2500
3000
UniversityofWashingtonCSEIntroductoryCourseAnnualEnrollment
(1-yearmovingtotal)
CSE143
CSE142
UWCSEintroductorycourseenrollment(1-yearrollingaverage)
0
100
200
300
400
500
600
700
800
900
1000
Top10First-ChoiceMajorsofUWConfirmedIncomingFreshmen
BusinessAdministration
ComputerScience&Engineering
Biology
MechanicalEngineering
Bioengineering
Psychology
Biochemistry
Aeronautics&Astronautics
Mathematics
Chemistry
Top10first-choicemajorsofUWconfirmedincomingfreshmen
2007-08EconomicsGovernmentSocialStudiesPsychologyEnglish&AmericanLiterature &LanguageHistoryAnthropologyHistory&LiteratureBiochemical SciencesAppliedMathematicsMolecular&Cellular BiologyHumanEvolutionaryBiologyNeurobiologyBiologyMathematicsSociologyChemistryPhysicsVisual&Environmental StudiesHistory&ScienceComputer ScienceEngineering&AppliedScience(AB)Chemical&PhysicalBiologyEnvironmental Science&PublicPolicyFineArts/History ofArt&Architecture
Top25concentrationsatHarvard2015-16EconomicsGovernmentComputer ScienceAppliedMathematicsPsychologySocialStudiesNeurobiologyStatisticsHumanDevelopmental&RegenerativeBiologyEnglishHistorySociologyHistory&LiteratureIntegrativeBiologyMolecular&Cellular BiologyMathematicsPhysicsChemistryHumanEvolutionaryBiologyHistory&ScienceEngineering&AppliedScience(SB)Biomedical EngineeringAnthropologyPhilosophyVisual&Environmental Studies Credit: HarryLewis
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
Life
Physical
Computational/Mathematical
DistributionofsciencemajorsatHarvard
Credit: HarryLewis
0
50
100
150
200
250
300
350
400
450
500
L&S:Undeclared
L&S:Math&PhysSciDiv
L&S:SocialSciencesDiv
L&S:UndergraduateDiv
L&S:Arts&HumanitiesDiv
L&S:BioSciencesDiv
L&S:Admin-CS
Berkeleyupper-divisionCSenrollmentfromL&SoutsideofCS
K-12too:CSAPparticipation,whilestillpathetic,isnowgrowing
Code.orgwillcauseallofthistoaccelerate
Hadi PartoviCode.org
Institutionalresponse:K-12
1983
IBMPCXT4.77MHz8088128KBRAMPCDOS2.0
401pagereport15pageindex
Scrollforward30years,to2013
WhichbegattheNextGenerationScienceStandards…
• In3outof4 highschoolsnationwide,computersciencethatincludesprogrammingisnotoffered– Butthat’sfarbetterthanjustayear
ortwoago!
• In22ofthe50states,computersciencedoesnotcounttowardsthemathorsciencegraduationrequirement– Butthat’sfarbetterthanjustayear
ortwoago!
Pathetic… butdespiteall,progressisbeingmade
Credit: Code.org
• Manypositivesigns– “SchoolsofComputerScience”areproliferating– Whatevertheprefix,thekeything isthatComputerSciencebeviewedasa
unitoftheentireuniversity• Computerscienceprogramsneedtoact thisway
• However
Institutionalresponse:highereducation
0"
10,000"
20,000"
30,000"
40,000"
50,000"
60,000"
70,000"
1966
"19
67"
1968
"19
69"
1970
"19
71"
1972
"19
73"
1974
"19
75"
1976
"19
77"
1978
"19
79"
1980
"19
81"
1982
"19
83"
1984
"19
85"
1986
"19
87"
1988
"19
89"
1990
"19
91"
1992
"19
93"
1994
"19
95"
1996
"19
97"
1998
"19
99"
2000
"20
01"
2002
"20
03"
2004
"20
05"
2006
"20
07"
2008
"20
09"
2010
"20
11"
2012
"
Computer)Science)Bachelors)Degrees)Granted)
– Sometendencytoviewcurrentsituationasatransient
• E.g.,hirelecturers;usefaculty fromotherfields– Facilitiesareahugeproblem
• Mustaccommodatescale• Mustrespondtoevolvingnatureofthefield
0
50,000
100,000
150,000
200,000
250,000
300,000
ComputerScience Engineering LifeSciences(incl.agricultural)
SocialSciences(incl.psychology)
PhysicalSciences(incl.environmental)
MathematicalSciences
Annualjobsavailablevs.degreesgranted
Annual jobsavailable Annual Bachelorsdegrees AnnualMastersdegrees Annual Doctoral degrees
BLSjobprojection data:http://www.bls.gov/emp/ind-occ-matrix/occupation.xlsxS&E Indicatorsdegreedata:http://www.nsf.gov/statistics/2016/nsb20161/uploads/1/12/at02-01.xlsx
BLS job projection data: http://www.bls.gov/emp/ind-occ-matrix/occupation.xlsxS&E Indicators degree data: http://www.nsf.gov/statistics/2016/nsb20161/uploads/1/12/at02-01.xlsx
0
50,000
100,000
150,000
200,000
250,000
300,000
ComputerScience Engineering LifeSciences(incl.agricultural)
SocialSciences(incl.psychology)
PhysicalSciences(incl.environmental)
MathematicalSciences
Annualjobsavailablevs.degreesgranted
Annual jobsavailable Annual Bachelorsdegrees AnnualMastersdegrees Annual Doctoral degrees
BLSjobprojection data:http://www.bls.gov/emp/ind-occ-matrix/occupation.xlsxS&E Indicatorsdegreedata:http://www.nsf.gov/statistics/2016/nsb20161/uploads/1/12/at02-01.xlsx
BLS job projection data: http://www.bls.gov/emp/ind-occ-matrix/occupation.xlsxS&E Indicators degree data: http://www.nsf.gov/statistics/2016/nsb20161/uploads/1/12/at02-01.xlsx
Roomforgrowth:AnnualSTEMjobopenings(BLS) vs.degreesgranted(NSF)
Isthisagreattimeorwhat?
http://lazowska.cs.washington.edu/NU.pdf,pptx