連想概念辞書の実装による意味ネットワークと比喩理解システムへの応用
4
0
0
全文
(2) < *= >+ ,? - " # $. 8/ 90 #1 $ :2 ;3
(3)
(4) 4 5 6 7 8 ? 9 @ # $ A : ;
(5) %B &C +!D ,"E #F $ G H ?
(6) %6 I
(7) &' ' ( (+
(8) ) , J2 KW CX DY E* F+ G6 LI "+ M, N2 OZ
(9) " [ P , 2: Q* ?\ M: N" O , #? $] ^ _
(10) &` a b
(11) c d
(12) + e 'f
(13) gR S
(14) Th i U, +j k V
(15) l / ^ m " * n o + p ? q r + V W X Y " s t
(16) P , 2 :. Integrated-and-Fire. Implementation of Associative Concept Dictionary for Semantic Network Structure and its Application for Metaphor Understanding System. and Shun Ishizaki. Takuya Sakaguchi. Abstract: We can find several models to achieve metaphor understanding with computer, including few models available for existing data about concepts. In this study, we proposed a computational model for metaphor understanding using an associative concept dictionary, which is a real data describing concepts and their relations. We implemented the data into a neural network of Integrated-and-Fire neuron model to build a semantic network structure, which was applied to the computational model to construct a metaphor understanding system. We developed the model with considering asymmetry of influence of vehicle and topic for metaphor understanding, and evaluated the system mentioning to its capability to differentiate an original metaphor and an exchanged metaphor in which vehicle and topic were exchanged from an original.. b + < P N S T 6 a b S T " s A M N < 7 C.
(17) + , ! " - . / 0 1 # 2 $ 3 % 4 & ' 5 ( 6 ) 7 " .* ' B }S "T w H " a b S T H d " + & 9 8D 9 E :, ;F < = G > 7 4 < H ? I @ A J B G 5 C K L M N < S T ¤ < ¥n ¦ S §T ¦" M N S TC ¨ " ¡ S \ T S s T " ¢ £ H " a b + © ª 1 ~ « ¬ 7 U 3A V " K O 6 N 1 P < 7 Q ( ) 7 H C I O R W ' SX TY ,Z ® ¯ « ¬ ° D ± ² ³ M N C ' H " [ \ ] ( ^ _ ` G 6 E M , N F a b S T Q c e d f g a )b ´S µT 1c <d 7 , p A tH "P r N s < [ } \ " W] ^ ¶_ 5 P . ( [ \ ] ( · ¸ ¹ º »
(18) h F i j k l m n + o 4 p ' / 0 E , qF 7 "K W L N " 2 3 1 < 6 1 < + o 1 q . C D ¼X 1Y Z S' TU ,3 1 'a Bb S 6T c d P ½N ¾ HC " º º " a b S T º ' B C .
(19). . 1.. 7. [1]. [2]. [2]. Integrated-and-Fire. [3]. 300. 60000. 2.. 2.1. 2.2 Integrated and Fire. < a7 b Sw Tx c ] d ( 6 S H T " 6 r O s A y t z u \ 6 1 G < 7v <` {1 |Ë Ë1 Ë Ë, Ë Ë Ë' ËB Ë} Ë" ~ Ë Ë q a uvwxyzyz{ | } ~ 4 5 . À Á 1 g
(20) E,FH ¤ s ¿ Â. D E,F' " g
(21) A"Â Ç ± È Ã É & Ã ' ¿ { Ê 4 Ä Å CÆ Integrated-and-Fire i. Graduate School of Media and Governance, Keio University. EPSP. t. ui(t). ui (t ) = ∑η (t − ti( f ) ) + ∑ ∑ wij ε (t − t (j f ) ) (1). - 1 −21−. f. j. f.
(22) Î " H g
(23) 6 Â H 56P g
(24) & Â Ï ÌÐ Í¿ !Ê Ã s ÑÒ¿ À 4 CÑÒÓ " 4Ô ÙH O iN FÕ 'N BÖ × CØ 6 % 1 Ö ×Ö ØÚ y 1( < Hp "5 w P x X Y i Z j ' k H l Û Ü m 1 ¢ C £ A E , F A ` G Ý Þ 6 1 < HÁ " ß å N ] ( Ý k \ æ ç à è * 1 é G G á â ê ã 7 Eä , Fä 'í Bî ' pB " ¿ p Ê ¤ A ë ï s ð ì P â N ì C ñ " ò 's Kì Lâ 43 1 < c 6 ~ E c, F ó' H "ô W \ + s * ì 1 â C 2. Q Xe Yf Zg '
(25) H h F i j k l m n E , aF b S T c, dF + o ú 1 7 " / õ 0 ö i + j * k 1 l C m û ÷ H\ "ø Où E +oüý þ'BC wij. j. ti(f). tj(f). i. i. j. f. EPSP. wij wji. wij = w ji = max(a / d ij + b, a / d ji + b) (2) a. b. a=1.0 b=0.30. Integrated-and-Fire. 3.. 3. 3.1. 2.3. Integrated-and-Fire. [4]. S T yó z Hô 6 . -A A" + w ' B M N < ; 7 " X Y Z ' H + 3 ]_ Á 1 < 6 1 C H % 1]" " H" Í !6 ; D 'C " W W ' a b S "T S T 4 ; 9 " O N6 Õ N§ ¦ gM
(26) N hFijkl m e§&<+o1 C
(27) g K
(28) L Eh ,F Fi 'j Hk " lN ma nb S +T oc Md 3 7 4< C û H' " n ^ ' g
(29) h¢ F£ i j k l1 " m ^ '
(30) å ! " q . C X å " 4 6 "" #a b H S T c% d$ 6 7+ o. 1 % g
(31) ¶ 6h 1F i ÷j \k 'l ( Am Ä& 'W S} "T ) 6 ] S T ¼ ( & N Õ N Sa Tb S T A c ¿ d Ê 4 C S T% "1 " H ¡\ST " HO # § ¦ M ST$N a # b ST6 ST$1 " X *' H # O N Õ SN T$ 6 + ", g
(32) C X Y h Z F ' i H j W k l . m ( n' ST . ²/7<014E,F L1 C . ST6 STA å " M N 6 " g
(33) \' h( F i j k < l O N m Õ Ä N ' W Í N ! P M A N ¿ Ê ¡\S 1 "÷ TA % ST 2 ¿ 7 Ê ÷ Ã\ 4 C 1 % " 1 W ST " 2. 1. GUI. 2.
(34)
(35) . . .
(36) $ % " & #! $ $ % % ) ' * + ( ( $% ,(-( $$$ %%% 02 . 1 3 / 20 ( 13 $$ %% 4 5 6 - ( .
(37)
(38)
(39)
(40) .
(41) . . . . . . .
(42)
(43)
(44)
(45) .
(46) .
(47) . . .
(48)
(49) .
(50) .
(51)
(52)
(53)
(54) . .
(55) .
(56) . . . . .
(57) .
(58) .
(59)
(60) . . .
(61)
(62) . . A. 0.474. 0.362. . U. I. B. B. C. . A B. C.
(63) . Target. :. 1. Source. Goal. Goal. 7j k l Sm TÄ g
(64) a b S T " g
(65) ' h × F M i " ÷ \ g
(66) Ñ Â * 1 A ¿ C Ê w 1 S T s < S G T 7 < H "
(67) × 1 4 C g W
(68) W ' Â H Ý s " S T s Í ! Î 66 × 41 ÍC ¨ ! Î " a b HS ´T c d ' yH zS T Bs 1. dij. wij. G. A. . ] H " ¨ a b S T c d { | M N < 1. :. [5]. 3.2 Goal Target. Source. (1). Target. - 2 −22−. t=0. Source.
(69)
(70) H31 4 P ' ¿ Ê M C
(71) . × ' ¿B Ê C W Ã W ' 5 7
(72) 6 6 H"H 7
(73) Ì c 6 ~ cj k å% N$ 8 1 #Â ; j 9 k 1 H " r % $¤: #rA H Â ;P " ! .
(74)
(75) N, F 'C ä W N < H Á w 1 x " ] = ( S T 6 ' " X E ))) *** # +++ $ ,,, % --- ... &/ 0 ' 1 ( Sô T 4 ¿ pÊ ' Ã × 15 C< 4 ( Ò @> %? " ó 7 6 7 24 35 + <P "" ÍÔ A ] " Ô S T 6 6 w õ S T ¨H ¡ \5 * ; 9 8< : SW T ÷ H\ " 1 y B 4 >$ ?5 " P C ` G C] ,7 0 - 1 .+ D ;¡ P\ NS T H " % 6 "Í A )n w F E 1 4 Ä Å CÆ > , - ? . 0 =55 1 È A Ç ± 6 È 4 É p ' " G ! w ¿ ÷ Ê \ 4 ' ( A " W 777 B C +++ /A @%& @ ' H ¿ ÊS 1T ( ) ( }5 "P Í A ] H & I S S T T 6 " H D2I E3J K L F M G 5 G ¨X }E , F ' J H W S T $ K A ¿ Ê < 4 P C N J H RI OO MJ K P NH Q! "4S T C ¨ " M D J S T SA T¿ 6L LM (N 1 < !! US == VT 00P 11Q H< !"4 " Ä Å Æ C c ä6 ~ê c5 JS T S TN O"6P 11 " XZ W Y[ F\ HG ]^ @_ @ 6 1 < !"4 H " ^ _b `c a \ 5 H -. 2 . ä ' B C de 5 `` fa H5
(76) 9':( & g & qK L E C, + F + U o 3 1 4 H " a Q b RS +T c h j0 i (1 B 8 C d . @{ "| JM N S T A S U T n S R õ V 1 T K ! Ò 5 " # $ % , BR - CM . P 6/ 11 M CN Ñ " 6 U 7 3
(77) 1 S T " á 6 W 7 B C + & R4 ' M5 ( k J !"ÍA X4C SHT t=2 ui ( 4) = ∑ ∑ wijε (t − t (j f ) ) = wiT ε (3) + wiSε (1) (3) j. f. Target. Source. (3). (3)=0.15. (1)=0.37. Target. Source. 5:12. SD. 1:2. [6]. (3). ui(t). =0.50. Target. Source. Goal. Ortony. [5]. 4.. 4.1. 5. 71. 142. 1. 1:. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71. - 3 −23−. A0%1@ l m. H I 2 3. . .
(78) #Y Z 7$ #5 7$ ¤ w x ] ( ¡ \ S T [ \ q A B< 7ä 6 ¹ " ] C 6 %1 1< WÛ N P( w x ]_ ^a _M ¡b \1 S< T1 H¨ " Û < 7' B 6 6 ` P N O C #5 : #k7$ h e c # j k ÙH d Ù% $$ 66 7.7. J S TH AS ¿T L' MB N < } " W A " ! H õ }û Ý f () ¡Ñ \Ò 1_ a äM 'G B7< 6ä ¹h D 1 .< C7)g n " ' H " ¡ \ S T' 7· )i ? @w ABx \ 6k³1"E , F TK j ÒNC . 1< "7
(79) Ì × l A m ] (QP N R CI9 n 6| or%$ 67. .# rHÂ ' H; j <k A=%$ 6#(} Â "; j k H HWP N N P . "X' Z pAp ¨ 17 CÑ L ' H]. © Cr 7 < W "qN P G 677 <
(80) " 6 / 1 4' Z S T 6 S T Ý Þ 5 P 7
(81) O 'Í ZA "7 s
(82) t § u 1 C 1< 1 v. - ] Q y1<D ¡ H"\ S T 6Í!1 S T w < 7 pA§ STA 5A3( 78Ò Cw }"xy y1 < H Í!D "Ý ¡ \ S T 6Î 7Í! { $#|` S } T $#~AQ5 $"wCV G 7<G 7<H #zH #¤ hS T A 4$#þCÑ j k "W $#kN P h S j T m A$ 1S Tv 6 "xy 'U 3 M N < 7 6"¡ \ Í!D " O X4C 4.2. 1. 71. Target. 21. Source. Target. Source. 50. 2. 21. 2:.
(83) # $+ % , - & . ! ' / ( ) " * * " * .$-0 * " " . ÝM N ¡ A"7\ S T
(84) 6Î )Í!1 S T '/ 1 'Z( 5 61"Í!Î < H "% 2±AYZ^_ < S T ; 4" M N S T 6 S T >8 9? " b 4A ;M N H "%pA æ P h N ý CW " ± 4Nr 1 : < Í!Î7A"` 6 a b S6T c dh . \ A BAÍ!Î §´µM N ( ) ( ;"=>4?@ABC (3). Goal. wij. 5:12. Target. Source. (2). a. b. EPSP. X * ' H "a b S T c d . Integrated-and-Fire. 5.. EN , F2 3 g1
(85) < h F i j k q l . D m E + , o F 1 " KW L " S T / ó 1 ô \ C O ¹ 1 " C O S TÍ A1 M" ¡ N \ S T 6 G ) 7 < 7 H y " zK L BE , SF T' 57 P
(86) / .¨) 4pA! C Target. Source.
(87) . [1] , , Computer Today, 2000-3, pp.34-39, 2000. [2] , , , , 8-4, pp.37-54, 2001. [3] W.Maass and C.M.Bishop, ”Pulsed Neural Networks”, MIT Press, 1999. [4] Collins,A.M. and Loftus,E.F., “A Spreading-Activation Theory of Semantic Processing”, Psychological Review, 82-6, pp.407-428, 1975. [5] Ortony,A., “Beyond literal similarity”, Psychological Review, 86, pp.161-180, 1979. [6] , : , , 65, pp.197-205, 1994b.. # $ % & ' ( ) * + , ! ". * - 7 . 8 / < = > ? , 0 1 @ 2 , 3 A 4 5B 6C 7 8 9 : ;. −24− - 4 - E.
(88)
関連したドキュメント
A knowledge of the basic definitions and results concerning locally compact Hausdorff spaces and continuous function spaces on them is required as well as some basic properties
We have described the classical loss network model similar to that of Kelly [9]. It also arises in variety of different contexts. Appropriate choices of A and C for the
This conjecture is not solved yet, and a good direction to solve it should be to build first a Quillen model structure on the category of weak ω-groupoids in the sense of
The excess travel cost dynamics serves as a more general framework than the rational behavior adjustment process for modeling the travelers’ dynamic route choice behavior in
Corollary 5 There exist infinitely many possibilities to extend the derivative x 0 , constructed in Section 9 on Q to all real numbers preserving the Leibnitz
Hence for a given multiscale network, we may perform many steps of model reduction until we obtain a reduced model which is simple enough to allow for extensive simulations, that
In the language of category theory, Stone’s representation theorem means that there is a duality between the category of Boolean algebras (with homomorphisms) and the category of
Keywords: continuous time random walk, Brownian motion, collision time, skew Young tableaux, tandem queue.. AMS 2000 Subject Classification: Primary:
