Share
Embed
Automatic Selection of Predicates for Common Sense Knowledge Expression
Ai Makabi, Hiroshi Matsumoto and Kazuhide Yamamoto. Automatic Selection of Predicates for Common Sense Knowledge Expression. Proceedings of the Conference of the Pacific Association for Computational Linguistics (PACLING 2013), no page numbers (2013.9)
自然言語処理研究室
September 04, 2013
Transcript
32 slides
-
Automa'c selec'on of predicates for common sense knowledge
expression Ai Makabi, Kazuhide Yamamoto, Hiroshi Matsumoto Nagaoka University of Technology -
D&(350$"-1 •! C$)1+A+,$/)&-)2-#+,,25+-#)($%/"#+0) –!E0&%%&'(&,)3-$4,+15+) –!F&05+)&%$"-#)$.)!"##"$%&'$&'%($")*'+,'%-./01% @*+0G*)H+%&03) I)&%)*"J+02-5).0$%))K$(<2G*)82'-5 H+*/$-*+) L$)B$")80$4*+)&) M+8)*2#+).$0)1$5)
#$)12*(2/,2-+N O! !"#$%)2*)&)-&%+)$.)&)&"'( O! PA+0($%+)82'-5)8+<&A2$0) 44WC$)#0&2-)&)1$5 +Q5Q)R$-A+0*&'$-&,)*B*#+% KR KR -
D&(350$"-1 •! C$)1+A+,$/)&-)2-#+,,25+-#)($%/"#+0) –!E0&%%&'(&,)3-$4,+15+) –!F&05+)&%$"-#)$.)!"##"$%&'$&'%($")*'+,'% I)&%)*"J+02-5).0$%))K$(<2G*)82'-5 L$)B$")80$4*+)&) M+8)*2#+).$0)1$5) #$)12*(2/,2-+N O!
!"#$%)2*)&)-&%+)$.)&"'( O! PA+0($%+)82'-5)8+<&A2$0) 44WC$)#0&2-)&)1$5 +Q5Q)R$-A+0*&'$-&,)*B*#+% KR KR ) ))S$("*)$-)%+#<$1*T) ))U)D"2,12-5)&)($%%$-)*+-*+)3-$4,+15+)8&*+)) ))))VRW:DX) ))U)K0$A212-5)&((+**28,+)0+/0+*+-#&'$-).$0)"*+)))) ))2-)-&#"0&,),&-5"&5+)/0$(+**2-5)#&*3*) ) -
Related Works 1/2 • Exis'ng Upper Ontologies (SUMO, Cyc, etc.)
– Contain many general concepts – e.g. Collec'on: book • A Type of: Informa'on bearing object the form of paper • Instance of: Kind of ar'fact not dis'nguished by brand or model • Merits: – Exploit rigorously-‐defined CSK • Demerits: – Knowledge representa'on cannot be matched fully with actual expressions -
Related Works 2/2 • Defineing the CSK as some rela'ons
are added to sentences/words (ConceptNet) – e.g. 犬(dog) • CapableOf: 散歩(walk), 寝る(sleep) • SymbolOf: 忠誠(loyalty), • Merits: – Defini'on is be_er suited to a natural language processing task • Demerits: – For the Japanese ConceptNet, the most concepts are collected manually • Coverage of CSK is excep'onally low -
E$&,)$.)#<+)W#"1B •! !"#$%&'(&,,B)($-*#0"(#)&)`&/&-+*+)RW:D)#<&#) (&-)8+)"',2;+1).$0)*+%&-'()&-&,B*2*)2-)-&#"0&,) ,&-5"&5+)/0$(+**2-5 W+#)$.)/0+12(&#+*) #<&#)($U$(("0)42#<)&))-$"-) a)RW:) ) 44T3A+08)
44T)&1]+('A+) 44T)A+08&,)-$"- A+08) 8&03) 0"- &1]+('A+) /0+_B) ("#+ A+08&,)-$"-) #$)#0&2-9)#$)80+&1 RW:)$.)b1$5c -
S2-&,)5$&,)U)PA+0A2+4)$.)#<+)RW:D (&# %+49)%+$4) #$)80+&1) /0+_B) #$)80+&1) /0+_B) &-2%&, R$%/"#+)&)*2%2,&02#B)8+#4++-)-$"-*) /"//B
B+,/) ))))))TTTT) ) 8&03) #$)80+&1) /0+_B) 1$5 R$%/&0+)&#) #<+)/0+12(&#+U,+A+, !550+5&#+))($-(+/#*)&*) &)"//+0)($-(+/#)b&-2%&,c) 8&*+1)$-)#<+)*2%2,&02#B R$-(+/#) V-$"-X &-2%&, &)"//+0)($-(+/#)b&-2%&,c) 8&*+1)$-)#<+)*2%2,&02#B RW:) V/0+12(&#+X -
Specific Property of CSK • We make the three hypothesis:
1) The predicate a is the CSK of the noun n when the pair of a and n are frequently co-‐occurred in sentences. 2) The predicate a which co-‐occurs with any nouns is not the appropriate CSK 3) Whether the predicate a is a correct CSK or not, it depends on the number of unique nouns which co-‐occurred with a. -
Specific Property of CSK • We make the three hypothesis:
1) The predicate a is the CSK of the noun n when the pair of a and n are frequently co-‐occurred in sentences. 2) The predicate a which co-‐occurs with any nouns is not the appropriate CSK 3) Whether the predicate a is a correct CSK or not, it depends on the number of unique nouns which co-‐occurred with a. -
W/+(2^()K0$/+0#B)$.)RW: •! M+)%&3+)#<+)#<0++)<B/$#<+*2*T) YX! C<+)/0+12(&#+()(2*)#<+)RW:)$.)#<+)-$"-(*(4<+-) #<+)/&20)$.)))&-1(*(&0+).0+d"+-#,B)($U$(("00+1)2-) *+-#+-(+*Q)) [X! C<+)/0+12(&#+()(4<2(<)($U$(("0*)42#<)&-B)-$"-*) 2*)-$#)#<+)&//0$/02&#+)RW:)
eX! M<+#<+0)#<+)/0+12(&#+)))2*)&(($00+(#)RW:)$0)-$#9) 2#)1+/+-1*)$-)#<+)-"%8+0)$.)"-2d"+)-$"-*) 4<2(<)($U$(("00+1)42#<))Q I)&_+-1)'*'#'$2345%&!6""*)+A+0B1&B) )/0+12(&#+)))))))))))$"7$% ($00+(#)RW:)$0)-$#9) 2#)1+/+-1*)$-)#<+)-"%8+0)$.)"-2d"+)-$"-*) )+A+0B1&B) ($U$(("0) 42#<)<25<) .0+d"+-(B C<+)/0+12(&#+ 2*)-$#)#<+)&//0$/02&#+)RW:) C<+)/0+12(&#+ b&_+-1c)2*)#<+)RW:) 42#<)<25<)/0$8&82,2#B -
!"#$%&'()*+,+('$-)$.)K0+12(&#+* C<+)#$/)Yf)/0+12(&#+*)&112-5)#$) &)-$"-)b )V+,+%+-#&0B)*(<$$,Xc 3V#$)+-0$,,)2-)*(<$$,X) *3V#$)+1"(&#+X) 5P3V8+X) GP3V8+($%+X) 3V#$)^-2*<)*(<$$,X)
3V#$)52A+),+**$-*X) 23V#$)#&3+)&-)+6&%X) :N73V&_+-1X) )3V#$),+&0-X) 3V#$)($&(<X)) <25< ,$4 C<+)/0+12(&#+*)/,&(+1)"//+0) 2-)#<+),2*#)&0+)($-*21+0+1) %$0+)&//0$/02&#+)&*)#<+)RW:) -
Specific Property of CSK • We make the three hypothesis:
1) The predicate a is the CSK of the noun n when the pair of a and n are frequently co-‐occurred in sentences. 2) The predicate a which co-‐occurs with any nouns is not the appropriate CSK 3) Whether the predicate a is a correct CSK or not, it depends on the number of unique nouns which co-‐occurred with a. -
C<+)#$/)Yf)/0+12(&#+*)&112-5)#$) &)-$"-)b )V+,+%+-#&0B)*(<$$,Xc K0+12(&#+*)42#<)<25<) ($U$(("00+-(+).0+d"+-(B) 42#<)&)-$"-)8"#)(&--$#) (<&0&(#+02;+)#<+)-$"- I-($00+(#)RW:) •! g+0*&',+)4$01*)
•! R$U$(("00+1)42#<)%&-B) -$"-* 3V#$)+-0$,,)2-)*(<$$,X) *3V#$)+1"(&#+X) 5P3V8+X) GP3V8+($%+X) 3V#$)^-2*<)*(<$$,X) 3V#$)52A+),+**$-*X) 23V#$)#&3+)&-)+6&%X) :N73V&_+-1X) )3V#$),+&0-X) 3V#$)($&(<X)) !"#$%&'()*+,+('$-)$.)K0+12(&#+* -
!" #!!" $!!!" $#!!" %!!!" %#!!" !" %!!" &!!" '!!"
(!!" $!!!" \%+05+-(+)12*#028"'$-)$.)/0+12(&#+*) 2-)#<+)#$/)Y9fff)-$"-*)) C<+)/0+12(&#+*)4<2(<).&,,)"-1+0)&)(+0#&2-)*($/+)($U$(("0)42#<) %&-B)-$"-*)a)L+,+#+)#<+)/0+12(&#+*)&*)+&,-,.*'(/0,&%#)1,23 ?"%8+0)$.)"-2d"+)-$"-*)($U$(("002-5)42#<)/0+12(&#+) ?"%8+0)$.)"-2d"+)/0+12(&#+*)) -
!" #!!" $!!!" $#!!" %!!!" %#!!" !" %!!" &!!" '!!"
(!!" $!!!" \%+05+-(+)12*#028"'$-)$.)/0+12(&#+*) 2-)#<+)#$/)Y9fff)-$"-*)) C<+)/0+12(&#+*)4<2(<).&,,)"-1+0)&)(+0#&2-)*($/+)($U$(("0)42#<) %&-B)-$"-*)a)L+,+#+)#<+)/0+12(&#+*)&*)+&,-,.*'(/0,&%#)1,23 ?"%8+0)$.)"-2d"+)-$"-*)($U$(("002-5)42#<)/0+12(&#+) ?"%8+0)$.)"-2d"+)/0+12(&#+*)) C<+)-"%8+0)$.)"-2d"+)/0+12(&#+*9) 4<2(<)($U$(("0)42#<)hff)-$"-*9)2*)Yfff -
!" #!!" $!!!" $#!!" %!!!" %#!!" !" %!!" &!!" '!!"
(!!" $!!!" C<+)/0+12(&#+*)4<2(<).&,,)"-1+0)&)(+0#&2-)*($/+)($U$(("0)42#<) %&-B)-$"-*)a)L+,+#+)#<+)/0+12(&#+*)&*)+&,-,.*'(/0,&%#)1,23 ?"%8+0)$.)"-2d"+)-$"-*)($U$(("002-5)42#<)/0+12(&#+) ?"%8+0)$.)"-2d"+)/0+12(&#+*)) \%+05+-(+)12*#028"'$-)$.)/0+12(&#+*) 2-)#<+)#$/)Y9fff)-$"-*)) ($U$(("002-5)42#<) %&-B)-$"-* ($U$(("002-5)42#<) .+4)-$"-* -
!" #!!" $!!!" $#!!" %!!!" %#!!" !" %!!" &!!" '!!"
(!!" $!!!" C<2*)($-#&2-*)#<+)2-($00+(#,B)/0+12(&#+*) 8&*+1)$-)<B/$#<+*2*)V[X) V*<&0/,B)2-(0+&*+1X C<+)/0+12(&#+*)4<2(<).&,,)"-1+0)&)(+0#&2-)*($/+)($U$(("0)42#<) %&-B)-$"-*)a)L+,+#+)#<+)/0+12(&#+*)&*)+&,-,.*'(/0,&%#)1,23 ?"%8+0)$.)"-2d"+)-$"-*)($U$(("002-5)42#<)/0+12(&#+) ?"%8+0)$.)"-2d"+)/0+12(&#+*)) \%+05+-(+)12*#028"'$-)$.)/0+12(&#+*) 2-)#<+)#$/)Y9fff)-$"-*)) -
Specific Property of CSK • We make the three hypothesis:
1) The predicate a is the CSK of the noun n when the pair of a and n are frequently co-‐occurred in sentences. 2) The predicate a which co-‐occurs with any nouns is not the appropriate CSK 3) Whether the predicate a is a correct CSK or not, it depends on the number of unique nouns which co-‐occurred with a. -
W/+(2^()K0$/+0#B)$.)RW: •! M+)%&3+)#<+)#<0++)<B/$#<+*2*T) YX! C<+)/0+12(&#+()(2*)#<+)RW:)$.)#<+)-$"-(*(4<+-) #<+)/&20)$.)))&-1(*(&0+).0+d"+-#,B)($U$(("00+1)2-) *+-#+-(+*Q)) [X! C<+)/0+12(&#+()(4<2(<)($U$(("0*)42#<)&-B)-$"-*) 2*)-$#)#<+)&//0$/02&#+)RW:)
eX! M<+#<+0)#<+)/0+12(&#+)))2*)&(($00+(#)RW:)$0)-$#9) 2#)1+/+-1*)$-)#<+)-"%8+0)$.)"-2d"+)-$"-*) 4<2(<)($U$(("00+1)42#<))Q W/+(2^()K0$/+0#B)$.)RW: M+)%&3+)#<+)#<0++)<B/$#<+*2*T) C<+)/0+12(&#+ #<+)/&20)$.) *+-#+-(+*Q)) C<+)/0+12(&#+ 2*)-$#)#<+)&//0$/02&#+)RW:) M<+#<+0)#<+)/0+12(&#+) W$0#) #<+)-$"-*) 8B)#<+) -"%8+0)$.) ($U$(("002-5) /0+12(&#+* 2-.$0%&'$-) /+0*$-) /0$1"(#) T) T) T) 0"--+0) 1&#&8&*+) /2&-$ W/+(2^()K0$/+0#B)$.)RW: 2*)#<+)RW:)$.)#<+)-$"-(*(4<+-) 2-.$0%&'$-) C<+)/0+12(&#+)$.)b0"-c) ($",1)-$#)(<&0&(#+02;+)#<+) -$"-)$.)b/+0*$-c &0+).0+d"+-#,B)($U$(("00+1)2-) 1&#&8&*+) C<+)/0+12(&#+)$.)b0"-c) ($",1)(<&0&(#+02;+)#<+) b0"--+0c C<+)-$"-)4<2(<)($U$(("0*)42#<)%&-B)/0+12(&#+*)(&-)-$#)8+) (<&0&(#+02;+1)8B)5+-+02()/0+12(&#+*9)<+-(+9)#<+)-"%8+0)$.)#<+20) 1+,+'-5)/0+12(&#+*)2*)%$0+)2-(0+&*+)#<&-)-$"-*)($U$(("002-5) 42#<)&).+4)/0+12(&#+*Q)) -
?"%8+0)$.)1+,+'-5)/0+12(&#+*) .$0)+&(<)-$"- Table I MBER OF DELETING PREDICATES FOR EACH
NOUN (N=THE UNIQUE NUMBER OF CO-OCCURRED PREDICATES) Scope of the nouns Deletion N≤700 427 700<N≤1,100 267 1,100<N≤1,600 143 1,600<N≤2,500 73 others 33 ver, the 33 predicates, which get deleted when can be used to nearly all nouns, so we consider are not common sense knowledge, and delete from A. Eval We c followin (1) D pr li (2) D pr li (3) R by no Deletion R$-*21+0)#<&#)#<+)ee) /0+12(&#+*)&0+)-$#)RW:9) &-1)1+,+#+).0$%)&,,) -$"-*)&*)2-($00+(#,B) /0+12(&#+*)) S:P3V"-1+0*#&-1X9)LD3V<&A+X9)KP3V*++9),$$3X9)GP3V8+($%+X9)) G63V-$#<2-5X9)FP3V#&3+9)&1$/#9)/0+.+0X9)E;P3V(&-X9)@P3V3-$4X9)) <P3V($%+X9)9L73V#<2-3X9)9963V%&-BX9)6P3V8+9)-++19)*<$$#X \6&%/,+)$.)1+,+'-5)/0+12(&#+* -
Added CSK for each noun he weighted scores for predicates
co-occurring with n sing Harman normalized frequency. A predicate is cor ommon sense knowledge for a noun when the predic core is high. The equation of Harman normalized freque s as follows (n: noun, a: predicate, na,n : appearance uency of predicate a with noun n). TF(a, n) = log2 (na,n + 1) log2 ( k nk,n) • The following equa'on computes weighted scores for predicates co-‐occurring with noun using Harman normalized frequency A predicate is appreciate as correct CSK for a noun when the predicate score is high. relate), Δ (do), ͔͚Δ (build, hang, run, lack) Figure 6. The deleting predicates for all noun se the selected predicates as common sense knowl- nd add them to each noun. In particular, we calculate ghted scores for predicates co-occurring with noun arman normalized frequency. A predicate is correct n sense knowledge for a noun when the predicate high. The equation of Harman normalized frequency llows (n: noun, a: predicate, na,n : appearance fre- of predicate a with noun n). T F (a, n) = log2 (na,n + 1) log2 ( k nk,n) (1) Figure 6. The deleting predicates for all noun use the selected predicates as common sense know nd add them to each noun. In particular, we calcula ighted scores for predicates co-occurring with nou Harman normalized frequency. A predicate is corre n sense knowledge for a noun when the predica high. The equation of Harman normalized frequen ollows (n: noun, a: predicate, na,n : appearance fr of predicate a with noun n). TF(a, n) = log2 (na,n + 1) log2 ( k nk,n) ( redicates for all noun es as common sense knowl- n. In particular, we calculate tes co-occurring with noun ency. A predicate is correct a noun when the predicate arman normalized frequency icate, na,n : appearance fre- n n). (na,n + 1) B. Evaluatio We take their assign follows (Tab The propose noun as the On the othe which frequ much higher “ݘ (dog)”, “Ұॹ (be to : noun : predicate : appearance frequency of predicate a with noun n -
Baselines 1) Do not delete the any predicates, just use
the weighted predicates by Harman normalized frequency (baseline 1) 2) Do not delete the any predicates, just use the weighted predicates by TF-‐IDF score (baseline 2) 3) Remove the 427 dele'ng predicates in N≤700, and use the weighted predicates by Harman normalized frequency (baseline 3) -
893#:*'%";%3&&<,$'+%:4'+<!32'&%3++<$,%2"%=+",> D&*+,2-+)Y D&*+,2-+)[ D&*+,2-+)e !//0$&(< :7U<&A+V :7U<&A+V SG63V1$)-$#) +&#X) 3V#$)#&3+)$"#)
.$0)&)4&,3X) GPU?'!"#'V %3-?'%2",'26'41% 1SG63V1$)-$#) 80++1X) @D=P3V802-5) "/X) 6PU?'V 3V#$),2A+X) :K>RA3V82#+)#$) 1+&#<X) #3V8+)*2(3X 5PU?'V . 3V#$)*&,+X) 8G63V1$)-$#) 8&03X) DQP3V#&3+) *$%+$-+)#$)#$4X) U#$),2A+V CI@63V."-X) +3V#$)52A+)&) ,+#<&,)2-]+('$-X) <OA3V,2A+X KPU*++V MA63V(<+&/X) '"3V#$)#+#<+0X ,&3V#$)#0&2-X G6U8+)-$-+V S:P3V"-1+0*#&-1X) ,&3V#$)#0&2-X) J8P3V8&03X) 67U*&BV $/3V#$)0+52*#+0X) MB>=P3V5+#)&,,) #<2-X) :S663V("#+X -
893#:*'%";%3&&<,$'+%:4'+<!32'&%3++<$,%2"%=+",> D&*+,2-+)Y D&*+,2-+)[ D&*+,2-+)e !//0$&(< :7U<&A+V :7U<&A+V SG63V1$)-$#) +&#X) 3V#$)#&3+)$"#)
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d"+-#,B)42#<)%&-B) -$"-* -
893#:*'%";%3&&<,$'+%:4'+<!32'&%3++<$,%2"%=+",> D&*+,2-+)Y D&*+,2-+)[ D&*+,2-+)e !//0$&(< :7U<&A+V :7U<&A+V SG63V1$)-$#) +&#X) 3V#$)#&3+)$"#)
.$0)&)4&,3X) GPU?'!"#'V %3-?'%2",'26'41% 1SG63V1$)-$#) 80++1X) @D=P3V802-5) "/X) 6PU?'V 3V#$),2A+X) :K>RA3V82#+)#$) 1+&#<X) #3V8+)*2(3X 5PU?'V . 3V#$)*&,+X) 8G63V1$)-$#) 8&03X) DQP3V#&3+) *$%+$-+)#$)#$4X) U#$),2A+V CI@63V."-X) +3V#$)52A+)&) ,+#<&,)2-]+('$-X) <OA3V,2A+X KPU*++V MA63V(<+&/X) '"3V#$)#+#<+0X ,&3V#$)#0&2-X G6U8+)-$-+V S:P3V"-1+0*#&-1X) ,&3V#$)#0&2-X) J8P3V8&03X) 67U*&BV $/3V#$)0+52*#+0X) MB>=P3V5+#)&,,) #<2-X) :S663V("#+X 6 5 U !//0$/02&#+) /0+12(&#+*)&0+)1+,+#+1 -
Error Analysis 1/3 • Although a predicate co-‐occurs with a
noun many 'mes, there are unrelated pairs – Do not check the dependency rela'on between them Solu'on: Use only the predicates which depend on the target nouns as candidate of CSK -
Error Analysis 2/3 • Could not assign nouns, which can
also be used as suffix to appropriate predicates – 美しい月です (This is the beau'ful moon) – 月ごとに決済する (We make a charge for each month) Solu'on: U'lize the rela'on of another co-‐occurred nouns e.g., If the “月” is co-‐occurred with a noun “太陽 (sun)”, it may mean the moon -
Error Analysis 3/3 • Include nouns which are used for
defining the rela'on of nouns – 原因 (cause) – 理由 (reason) Solu'on: Discuss how we limit the nouns of adding target -
Conclusion • Described the selec'on method of appropriate predicate
as CSK for construc'ng the CSKB. – Method for sta's'cally selec'ng CSK of nouns u'lizing the unique number of co-‐occurred predicates. • Evaluated sets of CSK which are assigned to each noun compared with three baselines – Demonstrated assumed characteris'cs of CKS in our study – Gave a subjec've evalua'on • Plan to make a quan'ta've evalua'on