crowddb answering queries with
play

CrowdDB: Answering queries with crowdsourcing ! Michael Franklin + , - PowerPoint PPT Presentation

UC Berkeley CrowdDB: Answering queries with crowdsourcing ! Michael Franklin + , Donald Kossmann*, Tim Kraska + Sukriti Ramesh*, and Reynold Xin + + UC Berkeley * ETH Zurich 1 ! !"#$%&'()*+ " ,-'%./)*+ # /#0!/+


  1. UC Berkeley CrowdDB: Answering queries with crowdsourcing ! Michael Franklin + , Donald Kossmann*, Tim Kraska + Sukriti Ramesh*, and Reynold Xin + + UC Berkeley * ETH Zurich 1

  2. ! !"#$%&'()*+ " ,-'%./)*+ # /#0!/+ 9#2,5:)+,00);+<=/2+0#%.&+%.+)#!48#.+)0,-/+ Algorithms Watson/IBM search Machines People 1//2+&/-'.%34/)+&#+25.,(%-,!!5+0%-6+7/)&+ #0/$,8."+0#%.&+ amplab 2

  3. 9'/+>?@+A,7+ Algorithms Watson/IBM search Machines People ?,6/+)/.)/+#B+2,&,+,&+)-,!/+75+8"'&!5+ %.&/"$,8."+,!"#$%&'()*+(,-'%./)*+,.2+0/#0!/+ 3 amplab

  4. C$#D2EF+ Algorithms Watson/IBM search Crowd DB Machines People ?,6/+)/.)/+#B+2,&,+,&+)-,!/+75+8"'&!5+ %.&/"$,8."+,!"#$%&'()*+(,-'%./)*+,.2+0/#0!/+ 4 amplab

  5. EFG',$2+H4/$%/)+ $%&'()*+,(&-. !//0-11. "(02-3.$('. I##"!/+ I##"!/0!/=*+?&.J+K%/D+C>+ LMNOF.+ P.&!J+F4)%./))+?,-'%./)+ >$(#.6*+1Q+ LMOOF.+ ?%-$#)#R+ S/2(#.2*+T>+ LMUOF.+ SELECT Market_Cap From Companies Where Company_Name = “IBM” Number of Rows: 0 Problem: Entity Resolution E amplab 5

  6. EFG',$2+H4/$%/)+ $%&'()*+,(&-. !//0-11. "(02-3.$('. I##"!/+ I##"!/0!/=*+?&.J+K%/D+C>+ LMNOF.+ P.&!J+F4)%./))+?,-'%./)+ >$(#.6*+1Q+ LMOOF.+ ?%-$#)#R+ S/2(#.2*+T>+ LMUOF.+ SELECT Market_Cap From Companies Where Company_Name = “Apple” Number of Rows: 0 Problem: Closed World Assumption C amplab 6

  7. EFG',$2+H4/$%/)+ SELECT Top_1(Image) From Pictures Where Theme = “Business Success” Number of Rows: 0 Problem: M Missing Intelligence amplab 7

  8. V,)5+H4/$%/)+ $%&'()*+,(&-. !//0-11. "(02-3.$('. I##"!/+ I##"!/0!/=*+?&.J+K%/D+C>+ LMNOF.+ P.&!J+F4)%./))+?,-'%./)+ >$(#.6*+1Q+ LMOOF.+ ?%-$#)#R+ S/2(#.2*+T>+ LMUOF.+ SELECT Market_Cap From Companies Where Company_Name = “IBM” $200Bn Number of Rows: 1 amplab 8

  9. @$/W5+V,)5+H4/$%/)+ $%&'()*+,(&-. !//0-11. "(02-3.$('. I##"!/+ I##"!/0!/=*+?&.J+K%/D+C>+ LMNOF.+ P.&!J+F4)%./))+?,-'%./)+ >$(#.6*+1Q+ LMOOF.+ ?%-$#)#R+ S/2(#.2*+T>+ LMUOF.+ SELECT Market_Cap From Companies Where Company_Name = “The Cool Software Company” $2xxBn Number of Rows: 1 amplab 9

  10. C$#D2)#4$-%."+ amplab 10

  11. ?%-$#&,)6%."+X+K%$&4,!%Y/2+Z4(,.)+ • " C4$$/.&+!/,2/$;+>(,Y#.+?/-',.%-,!+94$6+ • " S/34/)&#$)+0!,-/+Z4(,.+P.&/!!%"/.-/+9,)6)+[ZP9)\+ – " ?%.%(4(+0$%-/;+LOJON+ – " ]&'/$+0,$,(/&/$);+^#B+$/0!%-,)+[,))%".(/.&)\*+/=0%$,8#.*++ 41-0.5)3-06(7- *_ . – " >@PG7,)/2;+`-$/,&/Z%&[\a*+`"/&>))%".(/.&)[\a*+ `,00$#b/>))%".(/.&)[\a*+`B#$-/V=0%$/[\a+ – " S/34/)&#$)+,00$#b/+c#7)+,.2+0,5(/.&+ • " T#$6/$)+[,J6J,J+`&4$6/$)a\+-'##)/+c#7)*+2#+&'/(*+"/&+ 0,%2+ amplab 11

  12. C$#D2EF++ CrowdSQL Results 41-.38-.70%9/.3%.()19-0. :;<8(0/.=>-0?-1 ! CrowdDB Parser + MetaData T'/$/+&#+4)/+&'/+-$#D2;+ • " UI Optimizer UI Editor – " Z,b/+0/#0!/+<.2+2,&,+ Creation – " Z,b/+0/#0!/+2#+`B4YY5a+ -#(0,$%)#.)+ Executor UI Template Manager Statistics • " V34,!%&5+[',)'\+,.2+]$2/$%."+[)#$&\+ Files Access Methods HIT Management T'/$/+%&+2#/).:&+(,6/+)/.)/;+ • " – " Z,b/+0/#0!/+2#+34%-6+)#$&+#B+)/&)+ [%J/J*+94$6P&\+ – " >.5&'%."+&'/+-#(04&/$+,!$/,25+ Disk 1 2#/)+D/!!++ Disk 2 amplab 12

  13. C$#D2dHA+ee+dHAfff+ • " $@%1-/<A%0@/. ! ! .B'-)<A%0@/. – " dHA)+-!#)/2GD#$!2+,))4(08#.+%)+,+!%/+ – " P.g4/.-/)+34/$5+/=/-48#.+)&$,&/"%/)h#08#.)+ – " >&+&'/+(#(/.&*+D/+#.!5+,!!#D+,+$/)&$%-&/2+)/&+#B+34/$%/)+,",%.)&+ -$#D2G)#4$-/2+&,7!/)+ – " I#,!;+V=0!#$/+&'/+#0/.GD#$!2+,)+(4-'+,)+0#))%7!/+[/J"J*+,.)D/$+,+ 34/$5+,)+7/)&+,)+0#))%7!/+"%b/.+,+-/$&,%.+742"/&\++ • " $(78?)C. – " Vb/$5+$/)4!&+B$#(+&'/+-$#D2+%)+)&#$/2+ ! +9##+/=0/.)%b/+.#&+&#+2#+)#+ – " H4/$%/)+4)/+)&#$/2+$/)4!&)+D'/./b/$+0#))%7!/+ – " >.)D/$)+-',."/+7,)/2+#.+&'/+34/$5+'%)&#$5+,.2+-,-'/+7/',b%#$+ – " I#,!;+]i/$+(#$/+-#.&$#!+#b/$+-,-'%."*+99A*+_+ • " !)19-0.D>(@?3*. – " Z4(,.G%.04&G&#!/$,.&+34/$5+0$#-/))%."+ – " >&+&'/+(#(/.&+,+)%(0!/+)/&+#B+'/4$%)8-)++ – " I#,!;+-#.<2/.-/+%.&/$b,!)*+%&/$,8b/+%(0$#b/(/.&*_+ amplab 13

  14. C$#D2dHA+ ::E.FG3-)1?%)1 ; !+(9:.(1+0*:'0(31<2.''' !+(9:.(1+0*:'-,=3*.' CREATE C CROWD TABLE department ( CREATE TABLE company ( university STRING, name STRING PRIMARY KEY, department STRING, hq_address C CROWD STRING); phone_no STRING) PRIMARY KEY (university, department ) ; :"E.FG3-)1?%)1H.. !+(9:&;1,38''' !"#$%#"%&"'()*+,-(+.'/01++*2-34'5%6.78' SELECT * SELECT p FROM picture FROM companies WHERE subject = "Golden Gate Bridge" ~ “Big Blue” + WHERE Name ~ ORDER BY C CROWDORDER(p, "Which pic shows better %subject"); amplab 14

  15. j)/$+P.&/$B,-/+I/./$,8#.+ • " >+-!/,$+jP+%)+6/5+&#+$/)0#.)/+8(/+,.2+,.)D/$+ 34,!%&5J+ • " T/+-,.+!/b/$,"/+&'/+dHA+d-'/(,+&#+,4&#G "/./$,&/+jP++[/J"J*+]$,-!/+k#$()*+/&-J\+ amplab 15

  16. jP+D%&'+C#.&/=&+ amplab 16

  17. H4/$5+]08(%Y,8#.+,.2+V=/-48#.+ CREATE C CROWD TABLE department ( dVAVC9+l++ name STRING PRIMARY KEY kS]?+@S]kVdd]S+0*+EV@>S9?V19+2+ phone_no STRING); TZVSV+2J.,(/+e+0J2/0+ CREATE C CROWD TABLE professor ( >1E+0J.,(/+e`?%-',/!+mJ+C,$/5a+ name STRING PRIMARY KEY e-mail STRING dep STRING REF department(name) Please fi ll out the missing ); department data Department CS ! ! name="Carey" Phone p.dep=d.name MTJoin Submit (Dep) p.dep = d.name Department ! p.dep=d.name Please fi ll out the missing ! name= professor data MTProbe Name Carey Professor Department "Carey" (Professor) E-Mail name=Carey Department Professor Submit (b) Logical plan (c) Logical plan (d) Physical plan before optimization after optimization amplab 17

  18. H4/$5+]08(%Y,8#.+ Please fi ll out the missing department data CS Department ! name="Carey" ! Phone p.dep=d.name MTJoin Submit (Dep) p.dep = d.name Department ! p.dep=d.name Please fi ll out the missing ! name= professor data MTProbe Carey Name Professor Department "Carey" (Professor) E-Mail name=Carey Department Professor Submit (b) Logical plan (c) Logical plan (d) Physical plan before optimization after optimization I>@-.J(1-/.%'K&?L-0. @/$B#$(,.-/+P.)%"'n4!+H4/$5+A,."4,"/+[@PHA\+&/-'.%34/)++ • " &#+2/,!+D%&'+#0/.GD#$!2+,))4(08#.+ d%(0!/+)/&+#B+$4!/)+&#+0%-6+&'/+7/)&+0!,.+ • " d%(0!/+'/4$%)8-)+&#+)/&+&'/+-$#D2+0,$,(/&/$)+++ • " [/J"J*+$/0!%-,8#.+B,-&#$*+0$%-/+0/$+ZP9*+/&-J\+ amplab 18 +

  19. E/,!%."+D%&'+&'/+]0/.GT#$!2+ Focus Right Now: PK Queries 1 1 ! ! name="Carey" dVAVC9+l++ p.dep=d.name 1*1 kS]?+@S]kVdd]S+0*++ ! ! name= +++EV@>S9?V19+2+ Department ! TZVSV+0J2/0+e+2J.,(/+ p.dep=d.name ! "Carey" ! >1E+0J.,(/+e`C,$/5a+ 1 Professor Department + Professor 10 Soon dVAVC9+l++ 10 Stop After (10) kS]?+@S]kVdd]S+0*++ ! ! p.dep=d.name +++EV@>S9?V19+2+ ! p.dep=d.name 10 10*1 TZVSV+0J2/0+e+2J.,(/+ ! ! AP?P9+O*+NO++ Professor Department Professor Department ! Never dVAVC9+l++ ! p.dep=d.name kS]?+@S]kVdd]S+0*++ ! ! +++EV@>S9?V19+2+ Professor Department TZVSV+0J2/0+e+2J.,(/+ amplab 19

  20. H4/$5+]08(%Y,8#.+ Please fi ll out the missing department data CS Department ! name="Carey" ! Phone p.dep=d.name MTJoin Submit (Dep) p.dep = d.name Department ! p.dep=d.name Please fi ll out the missing ! name= professor data MTProbe Carey Name Professor Department "Carey" (Professor) E-Mail name=Carey Department Professor Submit (b) Logical plan (c) Logical plan (d) Physical plan before optimization after optimization • " C$/,&/)+4)/$+%.&/$B,-/+&/(0!,&/)+ • " d/!/-&+0'5)%-,!+#0/$,&#$)+ • " 1/D+H4/$5+]0/$,&#$);+ – " C$#D2+]0/$,&#$);+?9@$#7/*+?9m#%.*+?9k4.-8#.+ – " ]&'/$;+d9]@+>k9VS+[%J/J*+!%(%&\+ amplab 20 +

  21. ?9@$#7/+ • " d%(%!,$+&#+,+&,7!/G)-,.+D%&'+0$/2%-,&/+04)'G 2#D.)+ • " F,&-'/)+)/b/$,!+c#7)+%.&#+#./+ZP9+ • " P))4/)+,)+(,.5+$/34/)&)+%.+0,$,!!/!+,)+0#))%7!/++ "M#0%J- [@$#B/))#$\+ [7,)/2+#.+&'/+-,$2%.,!%&5+0$/2%-8#.\+ 1,(/eC,$/5+ • " E#/)+)%(0!/+34,!%&5+-#.&$#!+[34#$4(+b#&/)\+ • " `C,-'/)a+&'/+$/)4!&+%.)%2/+&'/+-#$$/)0#.2%."+ &,7!/+ ! +34/$%/)+',b/+)%2/G/i/-&)++ Please fi ll out the missing Please fi ll out the missing professor data Please fi ll out the missing professor data company data! Name Carey N ame Carey IBM Name E-Mail Department Headquarter CS Department name address Department E-Mail Phone Submit Submit Submit Crowd Column & Crowd Column & Denormalization Crowd Colums Crowd Colums 21 w/o foreign keys with foreign keys amplab

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend


More recommend