Unifying Synchronous Tree-Adjoining Grammars and Tree Transducers via Bimorphisms

Size: px
Start display at page:

Download "Unifying Synchronous Tree-Adjoining Grammars and Tree Transducers via Bimorphisms"

Transcription

1 Unifying ynchronous Tree-Adjoining Grmmrs nd Tree Trnsducers vi Bimorphisms turt M. hieer Division of Engineering nd Applied ciences Hrvrd University Cmridge, MA, UA turt M. hieer Unifying ynchronous Tree-Adjoining Grmmrs nd Tree Trnsducers vi Bimorphisms. In Proceedings of the 11th Conference of the Europen Chpter of the Assocition for Computtionl Linguistics (EACL-06), Trento, Itly, 3 7 April. Astrct We plce synchronous tree-djoining grmmrs nd tree trnsducers in the single overrching frmework of imorphisms, continuing the unifiction of synchronous grmmrs nd tree trnsducers initited y hieer (2004). Along the wy, we present new definition of the tree-djoining grmmr derivtion reltion sed on novel direct inter-reduction of TAG nd mondic mcro tree trnsducers. Tree trnsformtion systems such s tree trnsducers nd synchronous grmmrs hve seen renewed interest, sed on perceived relevnce to new pplictions, such s importing syntctic structure into sttisticl mchine trnsltion models or founding formlism for speech commnd nd control. The exct reltionship mong vriety of formlisms hs een uncler, with lrge numer of seemingly unrelted formlisms eing independently proposed or chrcterized. An initil step towrd unifying the formlisms ws tken (hieer, 2004) in mking use of the formllnguge-theoretic device of imorphisms, previously used to chrcterize the tree reltions definle y tree trnsducers. In prticulr, the tree reltions definle y synchronous tree-sustitution grmmrs (TG) were shown to e just those definle y liner complete imorphisms, therey providing for the first time cler reltionship etween synchronous grmmrs nd tree trnsducers. In this work, we show how the imorphism frmework cn e used to cpture more powerful formlism, synchronous tree-djoining grmmrs, providing further uniting of the vrious nd disprte formlisms. After some preliminries (ection 1), we egin y reclling the definition of tree-djoining grmmrs nd synchronous tree-djoining grmmrs (ection 2). We turn then to set of known results relting context-free lnguges, tree homomorphisms, tree utomt, nd tree trnsducers to extend them for the tree-djoining lnguges (ection 3), presenting these in terms of restricted kinds of functionl progrms over trees, using simple grmmticl nottion for descriing the progrms. This llows us to esily express generliztions of the notions: mondic mcro tree homomorphisms, utomt, nd trnsducers, which er (t lest some of) the sme interreltionships tht their trditionl simpler counterprts do (ection 4). Finlly, we use this chrcteriztion to plce the synchronous TAG formlism in the imorphism frmework (ection 5), further unifying tree trnsducers nd other synchronous grmmr formlisms. We lso, in pssing, provide new chrcteriztion of the reltion etween TAG derivtion nd derived trees, nd new simpler nd more direct proof of the equivlence of TALs nd the output lnguges of mondic mcro tree trnsducers. 1 Preliminries We will notte sequences with ngle rckets, e.g.,,,c, or where no confusion results, simply s c, with the empty string written ε. Trees will hve nodes leled with elements of RANKED ALPHABET, set of symols F, ech with non-negtive integer RANK or ARITY ssigned to it, determining the numer of children for nodes so leled. To emphsize the rity of symol, we will write it s prenthesized superscript, for instnce f (n) for symol f of rity n. Anlogously, we write F (n) for the set of symols in F with rity n. ymols with rity zero (F (0) ) re clled NULLARY symols or CON-

2 TANT. The set of nonconstnts is written F ( 1). To express incomplete trees, trees with holes witing to e filled, we will llow leves to e leled with vriles, in ddition to nullry symols. The set of TREE OVER A RANKED AL- PHABET F AND VARIABLE X, notted T(F,X), is the smllest set such tht (i) f T(F,X) for ll f F (0) ; (ii) x T(F,X) for ll x X; nd (iii) f (t 1,...,t n ) T(F,X) for ll f F ( 1), nd t 1,...,t n T(F,X). We revite T(F, /0), where the set of vriles is empty, s T(F), the set of GROUND TREE over F. We will lso mke use of the set of n numericlly ordered vriles X n = {x 1,...,x n }, nd write x, y, z s synonyms for x 1, x 2, x 3, respectively. Trees cn lso e viewed s mppings from TREE ADDREE, sequences of integers, to the lels of nodes t those ddresses. The ddress ε is the ddress of the root, 1 the ddress of the first child, 12 the ddress of the second child of the first child, nd so forth. We will use the nottion t/p to pick out the sutree of the node t ddress p in the tree t. Replcing the sutree of t t ddress p y tree t, written t[p t ] is defined s (using for the insertion of n element on list) t[ε t ] = t f (t 1,...,t n )[(i p) t ] = f (t 1,...,t i [p t ],...,t n ) for 1 i n. The HEIGHT of tree t, notted height(t), is defined s follows: height(x) = 0 for ll x X nd height( f (t 1,...,t n )) = 1+mx n i=1 height(t i) for ll f F. We cn use trees with vriles s CONTEXT in which to plce other trees. A tree in T(F,X n ) will e clled context, typiclly denoted with the symol C. For context C T(F,X n ) nd sequence of n trees t 1,...,t n T(F), the UBTITU- TION OF t 1,...,t n INTO C, notted C[t 1,...,t n ], is defined inductively s follows: ( f (u 1,...,u m ))[t 1,...,t n ] = f (u 1 [t 1,...,t n ],...,u m [t 1,...,t n ]) x i [t 1,...,t n ] = t i. A tree t T(F,X) is LINEAR if nd only if no vrile in X occurs more thn once in t. We will use nottion kin to BNF to specify equtions defining functionl progrms of vrious sorts. As n introduction to the nottion we will use, here is grmmr defining trees over rnked lphet nd vriles (essentilly identiclly to the definition given ove): f (n) F (n) x X ::= x 0 x 1 x 2 t T(F,X) ::= f (m) (t 1,...,t m ) x The nottion llows definition of clsses of expressions (e.g., F (n) ) nd specifies metvriles over them ( f (n) ). These clsses cn e primitive (F (n) ) or defined (X), even inductively in terms of other clsses or themselves (T(F,X)). We use the metvriles nd suscripted vrints on the right-hnd side to represent n ritrry element of the corresponding clss. Thus, the elements t 1,...,t m stnd for ritrry trees in T(F,X), nd x n ritrry vrile in X. Becuse numericlly suscripted versions of x pper explicitly on the right hnd side of the rule defining vriles, numericlly suscripted vriles (e.g., x 1 ) on the right-hnd side of ll rules re tken to refer to the specific elements of x, wheres otherwise suscripted elements (e.g., x i ) re tken genericlly. 2 Tree-Adjoining Grmmrs Tree djoining grmmr (TAG) is tree grmmr formlism distinguished y its use of tree djunction opertion. Trditionl presenttions of TAG, which we will ssume fmilirity with, tke the symols in elementry nd derived trees to e unrnked; nodes leled with given nonterminl symol my hve differing numers of children. (Joshi nd ches (1997) present good overview.) For exmple, foot nodes of uxiliry trees nd sustitution nodes hve no children, wheres the similrly leled root nodes must hve t lest one. imilrly, two nodes with the sme lel ut differing numers of children my mtch for the purpose of llowing n djunction (s the root nodes of α 1 nd β 1 in Figure 1). In order to integrte TAG with tree trnsducers, however, we move to rnked lphet, which presents some prolems nd opportunities. (In some wys, the rnked lphet definition of TAGs is slightly more elegnt thn the trditionl one.) Although the ulk of the lter discussion integrting TAGs nd trnsducers ssumes (without loss of expressivity (Joshi nd ches, 1997, fn. 6)) limited form of TAG tht includes djunction ut not sustitution, we define the more complete form here. We will thus tke the nodes of TAG trees to e leled with symols from rnked lphet F; given symol then hs fixed rity nd fixed

3 α 1 : α 2 : T β 1 : /0 β 2 : /0 α 1 : 1 β 1 : /0 β 2 : /0 ε : T c T 1 1 c Figure 1: mple TAG for the copy lnguge {wcw w {,} }. Figure 2: mple core-restricted TAG for the copy lnguge {wcw w {,} }. numer of children. However, in order to mintin informtion out which symols my mtch for the purpose of djunction nd sustitution, we tke the elements of F to e explicitly formed s pirs of n unrnked lel e nd n rity n. (For nottionl consistency, we will use e for unrnked nd f for rnked symols.) We will notte these elements, using nottion, s e (n), nd mke use of function to unrnk symols in F, so tht e (n) = e. To hndle foot nodes, for ech non-nullry symol e (i) F ( 1), we will ssocite new nullry symol e, which one cn tke to e the pir of e nd ; the set of such symols will e notted F. imilrly, for sustitution nodes, F will e the set of nullry symols e for ll e (i) F ( 1). These dditionl symols, since they re nullry, will necessrily pper only t the frontier of trees. Finlly, to llow null djoining constrints, for ech f F (i), we introduce symol f /0 lso of rity i, nd tke F /0 to e the set of ll such symols. We will extend the function to provide the unrnked symol ssocited with these symols s well, so e = e = e (i) /0 = e. A TAG is then qudruple F,,I,A, where F is rnked lphet; F is distinguished initil symol; I is the set of initil trees, finite suset of T(F F /0 F ); nd A is the set of uxiliry trees, finite suset of T(F F /0 F F ). An uxiliry tree β whose root is leled f must hve exctly one node leled with f F nd no other nodes leled in F ; this node is its foot node, its ddress notted foot(β). In Figure 1, α 1 nd α 2 re initil trees; β 1 nd β 2 re uxiliry trees. In order to llow reference to prticulr tree in the set P, we ssocite with ech tree in P unique index, conventionlly notted with suscripted α or β for initil nd uxiliry trees respectively. This further llows us to hve multiple instnces of tree in I or A, distinguished y their index. (We will use nottion y using the index nd the tree tht it nmes interchngly.) The trees re comined y two opertions, sustitution nd djunction. Under sustitution, node leled e (t ddress p) in tree α cn e replced y n initil tree α with the corresponding lel f t the root when f = e. The resulting tree, the sustitution of α t p in α, is α[p α ]. Under djunction, n internl node of α t p leled f F is split prt, replced y n uxiliry tree β rooted in f when f = f. The resulting tree, the djunction of β t p in α, is α[p β[foot(β) α/p]]. This definition (y requiring f to e in F, not F or F ) mintins the stndrd convention, without loss of expressivity, tht djunction is disllowed t foot nodes nd sustitution nodes. The TAG in Figure 1 genertes tree set whose yield is the non-context-free copy lnguge {wcw w {,} }. The rities of the nodes re suppressed, s they re cler from context. A derivtion tree D records the opertions over the elementry trees used to derive given derived tree. Ech node in the derivtion tree specifies n elementry tree α, the node s child sutrees D i recording the derivtions for trees tht re djoined or sustituted into tht tree. A method is required to record t which node in α the tree specified y child sutree D i opertes. For trees recording derivtions in context-free grmmrs, there re exctly s mny sustitution opertions s nonterminls on the right-hnd side of the rule used. Thus, child order in the derivtion tree cn e used to record the identity of the sustitution node. But for TAG trees, opertions occur throughout the tree, nd some, nmely djunctions, cn e optionl, so simple convention using child order is not possile. Trditionlly, the rnches in the derivtion tree hve een notted with the ddress of the node in the prent tree t which the child node opertes. Figure 4 presents derivtion tree () using this nottion, long with the corresponding derived tree () for the string c. For simplicity elow, we use stripped down TAG formlism, one tht loses no expressivity in wek genertive cpcity ut is esier for nlysis purposes. First, we mke ll djunction oligtory, in the

4 B /0 1 3 A B A 2 B cn define trnsducer s kind of functionl progrm, tht is, set of equtions chrcterized y the following grmmr for equtions Eqn. (The set of sttes is conventionlly notted Q, with memers notted q. One of the sttes is distinguished s the INITIAL TATE of the trnsducer.) 1 Figure 3: mple TAG tree mrked with dicritics to show the permuttion of operle nodes. sense tht if node in tree llows djunction, n djunction must occur there. To get the effect of optionl djunction, for instnce t node leled B, we dd vestigil tree of single node ε B = B, which hs no djunction sites nd does not itself modify ny tree tht it djoins into. It thus founds the recursive structure of derivtions. econd, now tht it is determinte whether n opertion must occur t node, the numer of children of node in derivtion tree is determined y the elementry tree t tht node; it is just the numer of djunction or sustitution nodes in the tree, the OPERABLE NODE. All tht is left to determine is the mpping etween child order in the derivtion tree nd node in the elementry tree leling the prent, tht is, permuttion π on the operle nodes (or equivlently, their ddresses), so tht the i-th child of node leled α in derivtion tree is tken to specify the tree tht opertes t the node π i in α. This permuttion cn e thought of s specified s prt of the elementry tree itself. For exmple, the tree in Figure 3, which requires opertions t the nodes t ddresses ε, 12, nd 2, my e ssocited with the permuttion 12,2,ε. This permuttion cn e mrked on the tree itself with numeric dicritics i, s shown in the figure. Finlly, s mentioned efore, we eliminte sustitution (Joshi nd ches, 1997, fn. 6). With these chnges, the smple TAG grmmr nd derivtion tree of Figures 1 nd 4() might e expressed with the core TAG grmmr nd derivtion tree of Figures 2 nd 4(c). 3 Tree Trnsducers, Homomorphisms, nd Automt 3.1 Tree Trnsducers Informlly, TREE TRANDUCER is function from T(F) to T(G) defined such tht the symol t the root ofthe input tree nd current stte determines n output context in which the recursive imges of the sutrees re plced. Formlly, we q Q f (n) F (n) g (n) G (n) x i X ::= x 0 x 1 x 2 Eqn ::= q( f (n) (x 1,...,x n )) = τ (n) τ (n) R (n) ::= g (m) (τ (n) 1,...,τ(n) m ) q j (x i ) where 1 i n Intuitively speking, the expressions in R (n) re right-hnd-side terms using vriles limited to the first n. For exmple, the grmmr llows definition of the following set of equtions defining tree trnsducer: 2 q( f (x)) = g(q (x),q(x)) q() = q ( f (x)) = f (q (x)) q () = This trnsducer llows for the following derivtion: q( f ( f ()) = g(q ( f (),q( f ()))) = g( f (q ()),g(q (),q())) = g( f (),g(,)) The reltion defined y tree trnsducer with initil stte q is { t,u q(t) = u}. By virtue of nondeterminism in the equtions, multiple equtions for given stte q nd symol f, tree trnsducers define true reltions rther thn merely functions. TREE HOMOMORPHIM re sutype of tree trnsducers, those with only single stte, hence essentilly stteless. Other sutypes of tree trnsducers cn e defined y restricting the trees τ 1 trictly speking, wht we define here re nondeterministic top-down tree trnsducers. 2 Full definitions of tree trnsducers typiclly descrie trnsducer in terms of set of sttes, n input nd output rnked lphet, nd n initil stte, in ddition to the set of trnsitions, tht is, defining equtions. We will leve off these detils, in the expecttion tht the sets of sttes nd symols cn e inferred from the equtions, nd the initil stte determined under convention tht it is the stte defined in the textully first eqution. Note lso tht we vil ourselves of consistent renming of the vriles x 1, x 2, nd so forth, where convenient for redility.

5 tht form the right-hnd sides of equtions, the elements of R (n) used. A trnsducer is LINEAR if ll such τ re liner; is COMPLETE if τ contins every vrile in X n ; is ε-free if τ X n ; is YMBOL-TO-YMBOL if height(τ) = 1; nd is DELABELING if τ is complete, liner, nd symolto-symol. Another sucse is TREE AUTOMATA, tree trnsducers tht compute prtil identity function; these re deleling tree trnsducers tht preserve the lel nd the order of rguments. Becuse they compute only the identity function, tree utomt re of interest for their domins, not the mppings they compute. Their domins define tree lnguges, in prticulr, the so-clled REGU- LAR TREE LANGUAGE. 3.2 The Bimorphism Chrcteriztion of Tree Trnsducers Tree trnsducers cn e chrcterized directly in terms of equtions defining simple kind of functionl progrm, s ove. There is n elegnt lterntive chrcteriztion of tree trnsducers in terms of constelltion of elements of the vrious sutypes of trnsducers homomorphisms nd utomt we hve introduced, clled imorphism. A imorphism is triple L,h i,h o, consisting of regulr tree lnguge L (or, equivlently, tree utomton) nd two tree homomorphisms h i nd h o. The tree reltion defined y imorphism is the set of tree pirs tht re generle from elements of the tree lnguge y the homomorphisms, tht is, L( L,h i,h o ) = { h i (t),h o (t) t L}. We cn limit ttention to imorphisms in which the input or output homomorphisms re restricted to certin type, liner (L), complete (C), epsilonfree (F), symol-to-symol (), deleling (D), or unrestricted (M). We will write B(I, O) where I nd O chrcterize suclss of homomorphisms for the set of imorphisms for which the input homomorphism is in the suclss indicted y I nd the output homomorphism is in the suclss indicted y O. Thus, B(D,M) is the set of imorphisms for which the input homomorphism is deleling ut the output homomorphism cn e ritrry. The tree reltions definle y tree trnsducers turn out to e exctly this clss B(D,M) (Comon et l., 1997). The imorphism notion thus llows us to chrcterize the tree trnsductions purely in terms of tree utomt nd tree homomorphisms. We hve shown (hieer, 2004) tht the tree reltions defined y synchronous tree-sustitution grmmrs were exctly the reltions B(LC, LC). Intuitively speking, the tree lnguge in such imorphism represents the set of derivtion trees for the synchronous grmmr, nd ech homomorphism represents the reltion etween the derivtion tree nd the derived tree for one of the projected tree-sustitution grmmrs. The homomorphisms re liner nd complete ecuse the tree reltion etween tree-sustitution grmmr derivtion tree nd its ssocited derived tree is exctly liner complete tree homomorphism. To chrcterize the tree reltions defined y synchronous tree-djoining grmmr, it similry suffices to find simple homomorphism-like chrcteriztion of the tree reltion etween TAG derivtion trees nd derived trees. In ection 5 elow, we show tht liner complete emedded tree homomorphisms, which we introduce next, serve this purpose. 4 Emedded Tree Trnsducers Emedded tree trnsducers re generliztion of tree trnsducers in which sttes re llowed to tke single dditionl rgument in restricted mnner. They correspond to restrictive sucse of mcro tree trnsducers with one recursion vrile. We use the term emedded tree trnsducer rther thn the more cumersome mondic mcro tree trnsducer for revity nd y nlogy with emedded pushdown utomt (ches nd Vijy-hnker, 1990), nother utomt-theoretic chrcteriztion of the tree-djoining lnguges. We modify the grmmr of trnsducer equtions to dd n extr rgument to ech occurrence of stte q. To highlight the specil nture of the extr rgument, it is written in ngle rckets efore the input tree rgument. We uniformly use the otherwise unused vrile x 0 for this rgument in the left-hnd side, nd dd x 0 s possile right-hnd side itself. Finlly, right-hnd-side occurrences of sttes my e pssed n ritrry further righthnd-side tree in this rgument. q Q f (n) F (n) x i X ::= x 0 x 1 x 2 Eqn ::= q [x 0 ] ( f (n) (x 1,...,x n )) = τ (n) τ (n) R (n) ::= f (m) (τ (n) 1,...,τ(n) m ) x 0 q j τ (n) j (x i ) where 1 i n

6 Emedded trnsducers re strictly more expressive thn trditionl trnsducers, ecuse the extr rgument llows unounded communiction etween positions unoundedly distnt in depth in the output tree. For exmple, simple emedded trnsducer cn compute the reversl of string, e.g., 1(2(2(nil))) reverses to 2(2(1(nil))). (This is not computle y trditionl tree trnsducer.) It is given y the following equtions: r (x) = r nil (x) r x 0 (nil) = x 0 r x 0 (1(x)) = r 1(x 0 ) (x) r x 0 (2(x)) = r 2(x 0 ) (x) (1) This is, of course, just the norml ccumulting reverse functionl progrm, expressed s n emedded trnsducer. The dditionl power of emedded trnsducers is, we will show in this section, exctly wht is needed to chrcterize the dditionl power tht TAGs represent over CFGs in descriing tree lnguges. In prticulr, we show tht the reltion etween TAG derivtion tree nd derived tree is chrcterized y deterministic liner complete emedded tree trnsducer (DL- CETT). The reltion etween tree-djoining lnguges nd emedded tree trnsducers my e implicit in series of previous results in the forml-lnguge theory literture. 3 For instnce, Fujiyoshi nd Ksi (2000) show tht liner, complete mondic context-free tree grmmrs generte exctly the tree-djoining lnguges vi norml form for spine grmmrs. eprtely, the reltion etween context-free tree grmmrs nd mcro tree trnsducers hs een descried, where the reltionship etween the mondic vrints of ech is implicit. Thus, tken together, n equivlence etween the tree-djoining lnguges nd the imge lnguges of mondic mcro tree trnsducers might e pieced together. In the present work, we define the reltion etween tree-djoining lnguges nd liner complete mondic tree trnsducers directly, simply, nd trnsprently, y giving explicit constructions in oth directions, crefully hndling the distinction etween the unrnked trees of tree-djoining grmmrs nd the rnked trees of mcro tree trnsducers nd other importnt issues of detil in the constructions. The proof requires reductions in oth directions. First, we show tht for ny TAG we cn construct DLCETT tht specifies the tree reltion etween the derivtion trees for the TAG nd the derived 3 We re indeted to Uwe Mönnich for this oservtion. trees. Then, we show tht for ny DLCETT we cn construct TAG such tht the tree reltion etween the derivtion trees nd derived trees is relted through simple homomorphism to the DL- CETT tree reltion. 4.1 From TAG to Trnsducer Given n elementry tree α with the lel A t its root, let the sequence π = π 1,...,π n e permuttion on the nodes in α t which djunction occurs. (We use this ordering y mens of the dicritic representtion elow.) Then, if α is n uxiliry tree, construct the eqution q A x 0 (α(x 1,...,x n )) = α nd if α is n initil tree, construct the eqution q A (α(x 1,...,x n )) = α where the right-hnd-side trnsformtion is defined y 4 A /0 (t 1,...,t n ) = A( t 1,..., t n ) k A(t 1,...,t n ) = q A A /0 (t 1,...,t n ) (x k ) A = x 0 = (2) Note tht the equtions re liner nd complete, ecuse ech vrile x i is generted once s the tree α is trversed, nmely t position π i in the trversl (mrked with i ), nd the vrile x 0 is generted t the foot node only. Thus, the generted emedded tree trnsducer is liner nd complete. Becuse only one eqution is generted per tree, the trnsducer is trivilly deterministic. By wy of exmple, we consider the core TAG grmmr given y the following trees: α : 1 A(e) β A : A /0 ( 1 B(), 2 C( 3 D(A ))) β B : 1 B(,B ) ε B : B ε C : C ε D : D 4 It my seem like trickery to use the dicritics in this wy, s they re not relly components of the tree eing trversed, ut merely reflexes of n extrinsic ordering. But their use is enign. The sme trnsformtion cn e defined, it more cumersomely, keeping the permuttion π seprte, y trcking the permuttion nd the current ddress p in revised trnsformtion π,p defined s follows: A /0 (t 1,...,t n ) π,p = A( t 1 π,p 1,..., t n π,p n ) A(t 1,...,t n ) π,p = q A A /0 (t 1,...,t n ) π,p (x π 1 (p) ) A π,p = x 0 π,p = We then use α π,ε for the trnsformtion of the tree α.

7 α 2 1 α 1 ε β 1 β 2 2 T c α 1 β 1 β 2 () () (c) Figure 4: Derivtion nd derived trees for the smple grmmrs: () derivtion tree for the grmmr of Figure 1; () corresponding derived tree; (c) corresponding derivtion tree for the core TAG version of the grmmr in Figure 2. trting with the uxiliry tree β A = A /0 ( 1 B(), 2 C( 3 D(A ))), the djunction sites, corresponding to the nodes leled B, C, nd D t ddresses 1, 2, nd 21, hve een ritrrily given preorder permuttion. We therefore construct the eqution s follows: q A x 0 (β A (x 1,x 2,x 3 )) = A /0 ( 1 B(), 2 C( 3 D(A ))) = A( 1 B(), 2 C( 3 D(A )) ) = A(q B B /0 () (x 1 ), 2 C( 3 D(A )) ) = A(q B B( ) (x 1 ), 2 C( 3 D(A )) ) = = A(q B B() (x 1 ),q C C(q D D(x 0 ) (x 3 )) (x 2 )) imilr derivtions for the remining trees yield the (deterministic liner complete) emedded tree trnsducer defined y the following set of equtions: q A (α(x 1 )) = q A A(e) (x 1 ) q A x 0 (β A (x 1,x 2,x 3 )) = A(q B B() (x 1 ),q C C(q D D(x 0 ) (x 3 )) (x 2 )) q B x 0 (β B (x 1 )) = q B B(,x 0 ) (x 1 ) q B x 0 (ε B ()) = x 0 q C x 0 (ε C ()) = x 0 q D x 0 (ε D ()) = x 0 We cn use this trnsducer to compute the derived tree for the derivtion tree α(β A (β B (ε B ),ε C,ε D )). q A (α(β A (β B (ε B ),ε C,ε D ))) = q A A(e) (β A (β B (ε B ),ε C,ε D )) = A( q B B() (β B (ε B )), q C C(q D D(A(e)) (ε D )) (ε C )) = A(q B B(,B()) (ε B ),C(q D D(A(e)) (ε D ))) = A(B(,B()),C(D(A(e)))) ε As finl step, useful lter for the imorphism chrcteriztion of synchronous TAG, it is strightforwrd to show tht the trnsducer so constructed is the composition of regulr tree lnguge nd liner complete emedded tree homomorphism. 4.2 From Trnsducer to TAG Given liner complete emedded tree trnsducer, we construct corresponding TAG s follows: For ech rule of the form q i [x 0 ] ( f (m) (x 1,...,x m )) = τ we uild tree nmed q i, f,τ. Where this tree ppers is determined solely y the stte q i, so we tke the root node of the tree to e the stte. Any foot node in the tree will lso need to e mrked with the sme lel, so we pss this informtion down s the tree is uilt inductively. The tree is therefore of the form q i /0( τ i ) where the right-hnd-side trnsformtion i constructs the reminder of the tree y the inductive wlk of τ, with the suscript noting tht the root is leled q i. f (t 1,...,t m ) i = f /0 ( t 1 i,..., t m i ) q j τ (x k ) i = k q j ( τ i ) x 0 i = q i i = Note tht t x 0, foot node is generted of the proper lel. (Becuse the eqution is liner, only one foot node is generted, nd it is leled ppropritely y construction.) Where recursive processing of the input tree occurs (q j τ (x l )), we generte tree tht dmits djunctions t q j. The role of the dicritic k is merely to specify the permuttion of operle nodes for interpreting derivtion trees; it sys tht the k-th child in derivtion tree rooted in the current elementry tree is tken to specify djunctions t this node. The trees generted y this TAG re intended to correspond to the outputs of the corresponding tree trnsducer. Becuse of the more severe constrints on TAG, in prticulr tht ll comintoril limittions on putting sutrees together must e mnifest in the lels in the trees themselves, the outputs ctully contin more structure thn the corresponding trnsducer output. In prticulr, the stte-leled nodes re merely for ookkeeping. A homomorphism removing these nodes gives the desired trnsducer output. Most importntly, then, the wek genertive cpcity of TAGs nd LCETTs re identicl.

8 ome exmples my clrify the construction. Recll the reversl emedded trnsducer in (1) ove. The construction ove genertes TAG contining the following trees. We hve given them indictive nmes rther thn the cumersome ones of the form q i, f,τ. α : r /0 (1 : r (nil)) β nil : r /0(r ) β 1 : r /0(1 : r (1 /0 (r ))) β 2 : r /0(1 : r (2 /0 (r ))) It is simple to verify tht the derivtion tree derives the tree α(β 1 (β 2 (β 2 (β nil )))) r(r 6 (2(r (2(r (1(r (nil)))))))) imple homomorphisms tht extrct the input function symols on the input nd drop the ookkeeping sttes on the output reduce these trees to 1(2(2(nil))) nd 2(2(1(nil))) respectively, just s for the corresponding tree trnsducer. 5 ynchronous TAGs s Bimorphisms The mjor dvntge of chrcterizing TAG derivtion in terms of tree trnsducers (vi the compiltion (2)) is the integrtion of synchronous TAGs into the imorphism frmework. A synchronous TAG (hieer, 1994) is composed of set of triples t L,t R, where the two trees t L nd t R re elementry trees nd is set of links specifying pirs of linked operle nodes from t L nd t R. Without loss of generlity, we cn stipulte tht ech operle node in ech tree is impinged upon y exctly one link in. (If node is unlinked, the triple cn never e used; if overlinked, set of replcement triples cn e multiplied out.) In this cse, projection of the triples on first or second component, with permuttion defined y the corresponding projections on the links, is exctly TAG s defined ove. Thus, derivtions proceed just s in single TAG except tht nodes linked y some link in re simultneously operted on y pired trees derived y the grmmr. In order to model synchronous grmmr formlism s imorphism, the well-formed derivtions of the synchronous formlism must e chrcterizle s regulr tree lnguge nd the reltion etween such derivtion trees nd ech of the pired derived trees s homomorphism of some sort. For synchronous tree-sustitution grmmrs, derivtion trees re regulr tree lnguges, nd the mp from derivtion to ech of the pired derived trees is liner complete tree homomorphism. Thus, synchronous tree-sustitution grmmrs fll in the clss of imorphisms B(LC,LC). The other direction cn e shown s well; ll imorphisms in B(LC, LC) define tree reltions expressile y n TG. A similr result follows immeditely for TAG. Crucilly relying on the result ove tht the derivtion reltion is DLCETT, we cn use the method of hieer (2004) directly to chrcterize the synchronous TAG tree reltions s just B(ELC, ELC). We hve thus integrted synchronous TAG with the other trnsducer nd synchronous grmmr formlisms flling under the imorphism umrell. Acknowledgements We wish to thnk Mrk Drs, Uwe Mönnich, Reecc Nesson, Jmes Rogers, nd Ken hn for helpful discussions on the topic of this pper. This work ws supported in prt y grnt II from the Ntionl cience Foundtion. References H. Comon, M. Duchet, R. Gilleron, F. Jcquemrd, D. Lugiez,. Tison, nd M. Tommsi Tree utomt techniques nd pplictions. Aville t: tt. Relese of Octoer 1, A. Fujiyoshi nd T. Ksi pinl-formed context-free tree grmmrs. Theory of Computing ystems, 33: Arvind Joshi nd Yves ches Treedjoining grmmrs. In G. Rozenerg nd A. lom, editors, Hndook of Forml Lnguges, volume 3, pges pringer, Berlin. Yves ches nd K. Vijy-hnker Deterministic left to right prsing of tree djoining lnguges. In Proceedings of the 28th Annul Meeting of the Assocition for Computtionl Linguistics, pges , Pittsurgh, Pennsylvni, 6 9 June. turt M. hieer Restricting the wekgenertive cpcity of synchronous tree-djoining grmmrs. Computtionl Intelligence, 10(4): , Novemer. Also ville s cmp-lg/ turt M. hieer ynchronous grmmrs s tree trnsducers. In Proceedings of the eventh Interntionl Workshop on Tree Adjoining Grmmr nd Relted Formlisms (TAG+7), pges 88 95, Vncouver, Cnd, My

9.3. Regular Languages

9.3. Regular Languages 9.3. REGULAR LANGUAGES 139 9.3. Regulr Lnguges 9.3.1. Properties of Regulr Lnguges. Recll tht regulr lnguge is the lnguge ssocited to regulr grmmr, i.e., grmmr G = (N, T, P, σ) in which every production

More information

A ppendix to. I soquants. Producing at Least Cost. Chapter

A ppendix to. I soquants. Producing at Least Cost. Chapter A ppendix to Chpter 0 Producing t est Cost This ppendix descries set of useful tools for studying firm s long-run production nd costs. The tools re isoqunts nd isocost lines. I soqunts FIGURE A0. SHOWS

More information

INF 4130 Exercise set 4

INF 4130 Exercise set 4 INF 4130 Exercise set 4 Exercise 1 List the order in which we extrct the nodes from the Live Set queue when we do redth first serch of the following grph (tree) with the Live Set implemented s LIFO queue.

More information

Outline. CSE 326: Data Structures. Priority Queues Leftist Heaps & Skew Heaps. Announcements. New Heap Operation: Merge

Outline. CSE 326: Data Structures. Priority Queues Leftist Heaps & Skew Heaps. Announcements. New Heap Operation: Merge CSE 26: Dt Structures Priority Queues Leftist Heps & Skew Heps Outline Announcements Leftist Heps & Skew Heps Reding: Weiss, Ch. 6 Hl Perkins Spring 2 Lectures 6 & 4//2 4//2 2 Announcements Written HW

More information

What is Monte Carlo Simulation? Monte Carlo Simulation

What is Monte Carlo Simulation? Monte Carlo Simulation Wht is Monte Crlo Simultion? Monte Crlo methods re widely used clss of computtionl lgorithms for simulting the ehvior of vrious physicl nd mthemticl systems, nd for other computtions. Monte Crlo lgorithm

More information

Cache CPI and DFAs and NFAs. CS230 Tutorial 10

Cache CPI and DFAs and NFAs. CS230 Tutorial 10 Cche CPI nd DFAs nd NFAs CS230 Tutoril 10 Multi-Level Cche: Clculting CPI When memory ccess is ttempted, wht re the possible results? ccess miss miss CPU L1 Cche L2 Cche Memory L1 cche hit L2 cche hit

More information

Pushdown Automata. Courtesy: Costas Busch RPI

Pushdown Automata. Courtesy: Costas Busch RPI Pushdown Automt Courtesy: Costs Busch RPI Pushdown Automt Pushdown Automt Pushdown Automt Pushdown Automt Pushdown Automt Pushdown Automt Non-Determinism:NPDA PDAs re non-deterministic: non-deterministic

More information

Roadmap of This Lecture

Roadmap of This Lecture Reltionl Model Rodmp of This Lecture Structure of Reltionl Dtbses Fundmentl Reltionl-Algebr-Opertions Additionl Reltionl-Algebr-Opertions Extended Reltionl-Algebr-Opertions Null Vlues Modifiction of the

More information

Today s Outline. One More Operation. Priority Queues. New Operation: Merge. Leftist Heaps. Priority Queues. Admin: Priority Queues

Today s Outline. One More Operation. Priority Queues. New Operation: Merge. Leftist Heaps. Priority Queues. Admin: Priority Queues Tody s Outline Priority Queues CSE Dt Structures & Algorithms Ruth Anderson Spring 4// Admin: HW # due this Thursdy / t :9pm Printouts due Fridy in lecture. Priority Queues Leftist Heps Skew Heps 4// One

More information

Measuring Search Trees

Measuring Search Trees Mesuring Serch Trees Christin Bessiere 1, Bruno Znuttini 2, nd Cèsr Fernández 3 1 LIRMM-CNRS, Montpellier, Frnce 2 GREYC, Cen, Frnce 3 Univ. de Lleid, Lleid, Spin Astrct. The SAT nd CSP communities mke

More information

THE FINAL PROOF SUPPORTING THE TURNOVER FORMULA.

THE FINAL PROOF SUPPORTING THE TURNOVER FORMULA. THE FINAL PROOF SUPPORTING THE TURNOVER FORMULA. I would like to thnk Aris for his mthemticl contriutions nd his swet which hs enled deeper understnding of the turnover formul to emerge. His contriution

More information

PSAS: Government transfers what you need to know

PSAS: Government transfers what you need to know PSAS: Government trnsfers wht you need to know Ferury 2018 Overview This summry will provide users with n understnding of the significnt recognition, presenttion nd disclosure requirements of the stndrd.

More information

Suffix Trees. Outline. Introduction Suffix Trees (ST) Building STs in linear time: Ukkonen s algorithm Applications of ST.

Suffix Trees. Outline. Introduction Suffix Trees (ST) Building STs in linear time: Ukkonen s algorithm Applications of ST. Suffi Trees Outline Introduction Suffi Trees (ST) Building STs in liner time: Ukkonen s lgorithm Applictions of ST 2 Introduction 3 Sustrings String is ny sequence of chrcters. Sustring of string S is

More information

CH 71 COMPLETING THE SQUARE INTRODUCTION FACTORING PERFECT SQUARE TRINOMIALS

CH 71 COMPLETING THE SQUARE INTRODUCTION FACTORING PERFECT SQUARE TRINOMIALS CH 7 COMPLETING THE SQUARE INTRODUCTION I t s now time to py our dues regrding the Qudrtic Formul. Wht, you my sk, does this men? It mens tht the formul ws merely given to you once or twice in this course,

More information

Drawing Finite State Machines in L A TEX and TikZ A Tutorial

Drawing Finite State Machines in L A TEX and TikZ A Tutorial Drwing Finite Stte Mchines in L A TEX nd TikZ A Tutoril Styki Sikdr nd Dvid Ching ssikdr@nd.edu Version 3 Jnury 7, 208 Introduction L A TEX (pronounced ly-tek) is n open-source, multipltform document preprtion

More information

Problem Set 2 Suggested Solutions

Problem Set 2 Suggested Solutions 4.472 Prolem Set 2 Suggested Solutions Reecc Zrutskie Question : First find the chnge in the cpitl stock, k, tht will occur when the OLG economy moves to the new stedy stte fter the government imposes

More information

Addition and Subtraction

Addition and Subtraction Addition nd Subtrction Nme: Dte: Definition: rtionl expression A rtionl expression is n lgebric expression in frction form, with polynomils in the numertor nd denomintor such tht t lest one vrible ppers

More information

Chapter 2: Relational Model. Chapter 2: Relational Model

Chapter 2: Relational Model. Chapter 2: Relational Model Chpter : Reltionl Model Dtbse System Concepts, 5 th Ed. See www.db-book.com for conditions on re-use Chpter : Reltionl Model Structure of Reltionl Dtbses Fundmentl Reltionl-Algebr-Opertions Additionl Reltionl-Algebr-Opertions

More information

3/1/2016. Intermediate Microeconomics W3211. Lecture 7: The Endowment Economy. Today s Aims. The Story So Far. An Endowment Economy.

3/1/2016. Intermediate Microeconomics W3211. Lecture 7: The Endowment Economy. Today s Aims. The Story So Far. An Endowment Economy. 1 Intermedite Microeconomics W3211 Lecture 7: The Endowment Economy Introduction Columbi University, Spring 2016 Mrk Den: mrk.den@columbi.edu 2 The Story So Fr. 3 Tody s Aims 4 Remember: the course hd

More information

Solutions to Exercises, Set 3

Solutions to Exercises, Set 3 Shool of Computer Siene, University of Nottinghm G5MAL Mhines nd their Lnguges, Spring 1 Thorsten Altenkirh Solutions to Exerises, Set 3 Fridy 3rd Mrh 1 1. () () L(+ +ǫ) = {L(E +F) = L(E) L(F)} L() L(

More information

MARKET POWER AND MISREPRESENTATION

MARKET POWER AND MISREPRESENTATION MARKET POWER AND MISREPRESENTATION MICROECONOMICS Principles nd Anlysis Frnk Cowell Note: the detil in slides mrked * cn only e seen if you run the slideshow July 2017 1 Introduction Presenttion concerns

More information

Controlling a population of identical MDP

Controlling a population of identical MDP Controlling popultion of identicl MDP Nthlie Bertrnd Inri Rennes ongoing work with Miheer Dewskr (CMI), Blise Genest (IRISA) nd Hugo Gimert (LBRI) Trends nd Chllenges in Quntittive Verifiction Mysore,

More information

A Fuzzy Inventory Model With Lot Size Dependent Carrying / Holding Cost

A Fuzzy Inventory Model With Lot Size Dependent Carrying / Holding Cost IOSR Journl of Mthemtics (IOSR-JM e-issn: 78-578,p-ISSN: 9-765X, Volume 7, Issue 6 (Sep. - Oct. 0, PP 06-0 www.iosrournls.org A Fuzzy Inventory Model With Lot Size Dependent Crrying / olding Cost P. Prvthi,

More information

(a) by substituting u = x + 10 and applying the result on page 869 on the text, (b) integrating by parts with u = ln(x + 10), dv = dx, v = x, and

(a) by substituting u = x + 10 and applying the result on page 869 on the text, (b) integrating by parts with u = ln(x + 10), dv = dx, v = x, and Supplementry Questions for HP Chpter 5. Derive the formul ln( + 0) d = ( + 0) ln( + 0) + C in three wys: () by substituting u = + 0 nd pplying the result on pge 869 on the tet, (b) integrting by prts with

More information

Roma December 2008 Cadastre and agriculture subsidies : the Spanish experience

Roma December 2008 Cadastre and agriculture subsidies : the Spanish experience Cdstre nd griculture susidies : the Spnish experience Evolution of the Common Agriculturl Policy (CAP), The role of Cdstre The LPIS nd the SIGPAC Towrds Interoperility Evolution of Common Agriculturl Policy

More information

USE OF MATHEMATICAL EXPRESSIONS CHALLENGES IN ECONOMICS CLASSROOMS. Ashita Raveendran, Lecturer, DESSH, NCERT, New Delhi

USE OF MATHEMATICAL EXPRESSIONS CHALLENGES IN ECONOMICS CLASSROOMS. Ashita Raveendran, Lecturer, DESSH, NCERT, New Delhi USE OF MATHEMATICAL EXPRESSIONS CHALLENGES IN ECONOMICS CLASSROOMS Ashit Rveendrn, Lecturer, DESSH, NCERT, New Delhi Mthemtics definition s Science of order (White hed (1929)) points on to its fetures

More information

Behavioural Differential Equations and Coinduction for Binary Trees

Behavioural Differential Equations and Coinduction for Binary Trees Behviourl Differentil Equtions nd Coinduction for Binry Trees Alexndr Silv nd Jn Rutten,2 Centrum voor Wiskunde en Informtic (CWI) 2 Vrije Universiteit Amsterdm (VUA) {ms,jnr}@cwi.nl Abstrct. We study

More information

Arithmetic and Geometric Sequences

Arithmetic and Geometric Sequences Arithmetic nd Geometric Sequences A sequence is list of numbers or objects, clled terms, in certin order. In n rithmetic sequence, the difference between one term nd the next is lwys the sme. This difference

More information

ASYMMETRIC SWITCHING COSTS CAN IMPROVE THE PREDICTIVE POWER OF SHY S MODEL

ASYMMETRIC SWITCHING COSTS CAN IMPROVE THE PREDICTIVE POWER OF SHY S MODEL Document de trvil 2012-14 / April 2012 ASYMMETRIC SWITCHIG COSTS CA IMPROVE THE PREDICTIVE POWER OF SHY S MODEL Evens Slies OFCE-Sciences-Po Asymmetric switching costs cn improve the predictive power of

More information

JOURNAL THE ERGODIC TM CANDLESTICK OSCILLATOR ROBERT KRAUSZ'S. Volume 1, Issue 7

JOURNAL THE ERGODIC TM CANDLESTICK OSCILLATOR ROBERT KRAUSZ'S. Volume 1, Issue 7 ROBERT KRUSZ'S JOURNL Volume 1, Issue 7 THE ERGODIC TM CNDLESTICK OSCILLTOR S ometimes we re lucky (due to our diligence) nd we find tool tht is useful nd does the jo etter thn previous tools, or nswers

More information

International Monopoly under Uncertainty

International Monopoly under Uncertainty Interntionl Monopoly under Uncertinty Henry Ary University of Grnd Astrct A domestic monopolistic firm hs the option to service foreign mrket through export or y setting up plnt in the host country under

More information

UNinhabited aerial vehicles (UAVs) are becoming increasingly

UNinhabited aerial vehicles (UAVs) are becoming increasingly A Process Algebr Genetic Algorithm Sertc Krmn Tl Shim Emilio Frzzoli Abstrct A genetic lgorithm tht utilizes process lgebr for coding of solution chromosomes nd for defining evolutionry bsed opertors is

More information

UNIT 7 SINGLE SAMPLING PLANS

UNIT 7 SINGLE SAMPLING PLANS UNIT 7 SINGLE SAMPLING PLANS Structure 7. Introduction Objectives 7. Single Smpling Pln 7.3 Operting Chrcteristics (OC) Curve 7.4 Producer s Risk nd Consumer s Risk 7.5 Averge Outgoing Qulity (AOQ) 7.6

More information

OPEN BUDGET QUESTIONNAIRE SOUTH AFRICA

OPEN BUDGET QUESTIONNAIRE SOUTH AFRICA Interntionl Budget Prtnership OPEN BUDGET QUESTIONNAIRE SOUTH AFRICA September 28, 2007 Interntionl Budget Prtnership Center on Budget nd Policy Priorities 820 First Street, NE Suite 510 Wshington, DC

More information

European Treaty Series - No. 124 EUROPEAN CONVENTION ON THE RECOGNITION OF THE LEGAL PERSONALITY OF INTERNATIONAL NON-GOVERNMENTAL ORGANISATIONS

European Treaty Series - No. 124 EUROPEAN CONVENTION ON THE RECOGNITION OF THE LEGAL PERSONALITY OF INTERNATIONAL NON-GOVERNMENTAL ORGANISATIONS Europen Trety Series - No. 124 EUROPEAN CONVENTION ON THE RECOGNITION OF THE LEGAL PERSONALITY OF INTERNATIONAL NON-GOVERNMENTAL ORGANISATIONS Strsourg, 24.IV.1986 2 ETS 124 Legl personlity of NGOs, 24.IV.1986

More information

Buckling of Stiffened Panels 1 overall buckling vs plate buckling PCCB Panel Collapse Combined Buckling

Buckling of Stiffened Panels 1 overall buckling vs plate buckling PCCB Panel Collapse Combined Buckling Buckling of Stiffened Pnels overll uckling vs plte uckling PCCB Pnel Collpse Comined Buckling Vrious estimtes hve een developed to determine the minimum size stiffener to insure the plte uckles while the

More information

The Market Approach to Valuing Businesses (Second Edition)

The Market Approach to Valuing Businesses (Second Edition) BV: Cse Anlysis Completed Trnsction & Guideline Public Comprble MARKET APPROACH The Mrket Approch to Vluing Businesses (Second Edition) Shnnon P. Prtt This mteril is reproduced from The Mrket Approch to

More information

Chapter55. Algebraic expansion and simplification

Chapter55. Algebraic expansion and simplification Chpter55 Algebric expnsion nd simplifiction Contents: A The distributive lw B The product ( + b)(c + d) C Difference of two squres D Perfect squres expnsion E Further expnsion F The binomil expnsion 88

More information

The Okun curve is non-linear

The Okun curve is non-linear Economics Letters 70 (00) 53 57 www.elsevier.com/ locte/ econbse The Okun curve is non-liner Mtti Viren * Deprtment of Economics, 004 University of Turku, Turku, Finlnd Received 5 My 999; ccepted 0 April

More information

Problem Set for Chapter 3: Simple Regression Analysis ECO382 Econometrics Queens College K.Matsuda

Problem Set for Chapter 3: Simple Regression Analysis ECO382 Econometrics Queens College K.Matsuda Problem Set for Chpter 3 Simple Regression Anlysis ECO382 Econometrics Queens College K.Mtsud Excel Assignments You re required to hnd in these Excel Assignments by the due Mtsud specifies. Legibility

More information

OPEN BUDGET QUESTIONNAIRE MACEDONIA

OPEN BUDGET QUESTIONNAIRE MACEDONIA Interntionl Budget Prtnership OPEN BUDGET QUESTIONNAIRE MACEDONIA September 28, 2007 Interntionl Budget Prtnership Center on Budget nd Policy Priorities 820 First Street, NE Suite 510 Wshington, DC 20002

More information

A Closer Look at Bond Risk: Duration

A Closer Look at Bond Risk: Duration W E B E X T E S I O 4C A Closer Look t Bond Risk: Durtion This Extension explins how to mnge the risk of bond portfolio using the concept of durtion. BOD RISK In our discussion of bond vlution in Chpter

More information

Math F412: Homework 4 Solutions February 20, κ I = s α κ α

Math F412: Homework 4 Solutions February 20, κ I = s α κ α All prts of this homework to be completed in Mple should be done in single worksheet. You cn submit either the worksheet by emil or printout of it with your homework. 1. Opre 1.4.1 Let α be not-necessrily

More information

Effects of Entry Restriction on Free Entry General Competitive Equilibrium. Mitsuo Takase

Effects of Entry Restriction on Free Entry General Competitive Equilibrium. Mitsuo Takase CAES Working Pper Series Effects of Entry Restriction on Free Entry Generl Competitive Euilirium Mitsuo Tkse Fculty of Economics Fukuok University WP-2018-006 Center for Advnced Economic Study Fukuok University

More information

Chapter 4. Profit and Bayesian Optimality

Chapter 4. Profit and Bayesian Optimality Chpter 4 Profit nd Byesin Optimlity In this chpter we consider the objective of profit. The objective of profit mximiztion dds significnt new chllenge over the previously considered objective of socil

More information

Inequality and the GB2 income distribution

Inequality and the GB2 income distribution Working Pper Series Inequlity nd the GB2 income distribution Stephen P. Jenkins ECINEQ WP 2007 73 ECINEC 2007-73 July 2007 www.ecineq.org Inequlity nd the GB2 income distribution Stephen P. Jenkins* University

More information

Burrows-Wheeler Transform and FM Index

Burrows-Wheeler Transform and FM Index Burrows-Wheeler Trnsform nd M Index Ben ngmed You re free to use these slides. If you do, plese sign the guestbook (www.lngmed-lb.org/teching-mterils), or emil me (ben.lngmed@gmil.com) nd tell me briefly

More information

JFE Online Appendix: The QUAD Method

JFE Online Appendix: The QUAD Method JFE Online Appendix: The QUAD Method Prt of the QUAD technique is the use of qudrture for numericl solution of option pricing problems. Andricopoulos et l. (00, 007 use qudrture s the only computtionl

More information

164 CHAPTER 2. VECTOR FUNCTIONS

164 CHAPTER 2. VECTOR FUNCTIONS 164 CHAPTER. VECTOR FUNCTIONS.4 Curvture.4.1 Definitions nd Exmples The notion of curvture mesures how shrply curve bends. We would expect the curvture to be 0 for stright line, to be very smll for curves

More information

The IndoDairy Smallholder Household Survey From Farm-to-Fact

The IndoDairy Smallholder Household Survey From Farm-to-Fact The Centre for Glol Food nd Resources The IndoDiry Smllholder Household Survey From Frm-to-Fct Fctsheet 7: Diry Frming Costs, Revenue nd Profitility Bckground This fctsheet uilds on the informtion summrised

More information

Chapter 02: International Flow of Funds

Chapter 02: International Flow of Funds Chpter 02: Interntionl Flow of Funds 1. Recently, the U.S. experienced n nnul lnce of trde representing.. lrge surplus (exceeding $100 illion). smll surplus c. level of zero d. deficit d 2. A high home

More information

OPEN BUDGET QUESTIONNAIRE FRANCE

OPEN BUDGET QUESTIONNAIRE FRANCE Interntionl Budget Project OPEN BUDGET QUESTIONNAIRE FRANCE October 2005 Interntionl Budget Project Center on Budget nd Policy Priorities 820 First Street, NE Suite 510 Wshington, DC 20002 www.interntionlbudget.org

More information

International Budget Partnership OPEN BUDGET QUESTIONNAIRE POLAND

International Budget Partnership OPEN BUDGET QUESTIONNAIRE POLAND Interntionl Budget Prtnership OPEN BUDGET QUESTIONNAIRE POLAND September 28, 2007 Interntionl Budget Prtnership Center on Budget nd Policy Priorities 820 First Street, NE Suite 510 Wshington, DC 20002

More information

NORTH YORKSHIRE PENSION FUND GOVERNANCE COMPLIANCE STATEMENT

NORTH YORKSHIRE PENSION FUND GOVERNANCE COMPLIANCE STATEMENT NORTH YORKSHIRE PENSION FUND GOVERNANCE COMPLIANCE STATEMENT TABLE OF CONTENTS Section Pge 1 INTRODUCTION 2 2 GOVERNANCE ARRANGEMENTS 2 3 REPRESENTATION AND MEETINGS 4 4 OPERATIONAL PROCEDRES 5 5 KEY POLICY

More information

UNIVERSITY OF NOTTINGHAM. Discussion Papers in Economics BERTRAND VS. COURNOT COMPETITION IN ASYMMETRIC DUOPOLY: THE ROLE OF LICENSING

UNIVERSITY OF NOTTINGHAM. Discussion Papers in Economics BERTRAND VS. COURNOT COMPETITION IN ASYMMETRIC DUOPOLY: THE ROLE OF LICENSING UNIVERSITY OF NOTTINGHAM Discussion Ppers in Economics Discussion Pper No. 0/0 BERTRAND VS. COURNOT COMPETITION IN ASYMMETRIC DUOPOLY: THE ROLE OF LICENSING by Arijit Mukherjee April 00 DP 0/0 ISSN 160-48

More information

Choice of strategic variables under relative profit maximization in asymmetric oligopoly

Choice of strategic variables under relative profit maximization in asymmetric oligopoly Economics nd Business Letters () 5-6 04 Choice of strtegic vriles under reltive profit mximiztion in symmetric oligopoly Atsuhiro Stoh Ysuhito Tnk * Fculty of Economics Doshish University Kyoto Jpn Received:

More information

1. Detailed information about the Appellant s and Respondent s personal information including mobile no. and -id are to be furnished.

1. Detailed information about the Appellant s and Respondent s personal information including mobile no. and  -id are to be furnished. Revised Form 36 nd Form 36A for filing ppel nd cross objection respectively before income tx ppellte tribunl (Notifiction No. 72 dted 23.10.2018) Bckground CBDT issued drft notifiction vide press relese

More information

On the Complexity of Computing the Justification Status of an Argument

On the Complexity of Computing the Justification Status of an Argument On the Complexity of Computing the Justifiction Sttus of n Argument dbi Reserch Seminr, Vienn Wolfgng Dvořák Institute of Informtion Systems, Vienn University of Technology Oct 13, 2011 Supported by the

More information

BEFORE THE ENVIRONMENT COURT

BEFORE THE ENVIRONMENT COURT BEFORE THE ENVIRONMENT COURT I MUA I TE KOOTI TAIAO O AOTEAROA ENV-2018-AKL-000078 Under the Resource Mngement Act 1991 (the Act) In the mtter of direct referrl of n ppliction for resource consent for

More information

The MA health reform and other issues

The MA health reform and other issues The MA helth reorm nd other issues Gruer: three key issues or ny reorm Poolin Need wy to ornize helth cre other thn need Otherwise -- dverse selection Prolem: current system leves out smll irms Aordility

More information

3: Inventory management

3: Inventory management INSE6300 Ji Yun Yu 3: Inventory mngement Concordi Februry 9, 2016 Supply chin mngement is bout mking sequence of decisions over sequence of time steps, fter mking observtions t ech of these time steps.

More information

OPEN BUDGET QUESTIONNAIRE

OPEN BUDGET QUESTIONNAIRE Interntionl Budget Project OPEN BUDGET QUESTIONNAIRE UNITED STATES October 2005 Interntionl Budget Project Center on Budget nd Policy Priorities 820 First Street, NE Suite 510 Wshington, DC 20002 www.interntionlbudget.org

More information

1 Manipulation for binary voters

1 Manipulation for binary voters STAT 206A: Soil Choie nd Networks Fll 2010 Mnipultion nd GS Theorem Otoer 21 Leturer: Elhnn Mossel Srie: Kristen Woyh In this leture we over mnipultion y single voter: whether single voter n lie out his

More information

Gridworld Values V* Gridworld: Q*

Gridworld Values V* Gridworld: Q* CS 188: Artificil Intelligence Mrkov Deciion Procee II Intructor: Dn Klein nd Pieter Abbeel --- Univerity of Cliforni, Berkeley [Thee lide were creted by Dn Klein nd Pieter Abbeel for CS188 Intro to AI

More information

Trigonometry - Activity 21 General Triangle Solution: Given three sides.

Trigonometry - Activity 21 General Triangle Solution: Given three sides. Nme: lss: p 43 Mths Helper Plus Resoure Set. opyright 003 rue. Vughn, Tehers hoie Softwre Trigonometry - tivity 1 Generl Tringle Solution: Given three sides. When the three side lengths '', '' nd '' of

More information

OPEN BUDGET QUESTIONNAIRE

OPEN BUDGET QUESTIONNAIRE Interntionl Budget Prtnership OPEN BUDGET QUESTIONNAIRE SOUTH KOREA September 28, 2007 Interntionl Budget Prtnership Center on Budget nd Policy Priorities 820 First Street, NE Suite 510 Wshington, DC 20002

More information

Reinforcement Learning. CS 188: Artificial Intelligence Fall Grid World. Markov Decision Processes. What is Markov about MDPs?

Reinforcement Learning. CS 188: Artificial Intelligence Fall Grid World. Markov Decision Processes. What is Markov about MDPs? CS 188: Artificil Intelligence Fll 2010 Lecture 9: MDP 9/2/2010 Reinforcement Lerning [DEMOS] Bic ide: Receive feedbck in the form of rewrd Agent utility i defined by the rewrd function Mut (lern to) ct

More information

A portfolio approach to the optimal funding of pensions

A portfolio approach to the optimal funding of pensions Economics Letters 69 (000) 01 06 www.elsevier.com/ locte/ econbse A portfolio pproch to the optiml funding of pensions Jysri Dutt, Sndeep Kpur *, J. Michel Orszg b, b Fculty of Economics University of

More information

Chapter 3: The Reinforcement Learning Problem. The Agent'Environment Interface. Getting the Degree of Abstraction Right. The Agent Learns a Policy

Chapter 3: The Reinforcement Learning Problem. The Agent'Environment Interface. Getting the Degree of Abstraction Right. The Agent Learns a Policy Chpter 3: The Reinforcement Lerning Problem The Agent'Environment Interfce Objectives of this chpter: describe the RL problem we will be studying for the reminder of the course present idelized form of

More information

Get Solution of These Packages & Learn by Video Tutorials on KEY CONCEPTS

Get Solution of These Packages & Learn by Video Tutorials on  KEY CONCEPTS FREE Downlod Study Pckge from wesite: www.tekoclsses.com & www.mthsbysuhg.com Get Solution of These Pckges & Lern y Video Tutorils on www.mthsbysuhg.com KEY CONCEPTS THINGS TO REMEMBER :. The re ounded

More information

Autarky more likely than complete specialization

Autarky more likely than complete specialization Autrky more likely thn complete specilition Antonio Quesd Deprtment d Economi, Universitt Rovir i Virgili, Avingud de l Universitt, 4304 Reus, Spin 3th Jnury 00 38.9 Astrct This pper mesures the strength

More information

"Multilateralism, Regionalism, and the Sustainability of 'Natural' Trading Blocs"

Multilateralism, Regionalism, and the Sustainability of 'Natural' Trading Blocs "Multilterlism, Regionlism, nd the Sustinility of 'Nturl' Trding Blocs" y Eric Bond Deprtment of Economics Penn Stte June, 1999 Astrct: This pper compres the mximum level of world welfre ttinle in n incentive

More information

Tble 1: Syntx of LOTOS/T+ E :: stop (untimed dedlock) j exit (successful termintion) j ; E (ction prex, untimed) j [P (t; x)]; E (ction prex, timed) j

Tble 1: Syntx of LOTOS/T+ E :: stop (untimed dedlock) j exit (successful termintion) j ; E (ction prex, untimed) j [P (t; x)]; E (ction prex, timed) j Protocol Synthesis from Timed nd Structured Specictions Akio Nkt Teruo Higshino Kenichi Tniguchi Dept. of Informtion nd Computer Sciences, Osk University Toyonk, Osk 56, Jpn Abstrct In this pper, we propose

More information

checks are tax current income.

checks are tax current income. Humn Short Term Disbility Pln Wht is Disbility Insurnce? An esy explntion is; Disbility Insurnce is protection for your pycheck. Imgine if you were suddenly disbled, unble to work, due to n ccident or

More information

COMPLETE COVER - `ADD ON' TO TRAVEL INSURANCE POLICIES CONFERENCE & EVENT ORGANISERS SCHEDULE

COMPLETE COVER - `ADD ON' TO TRAVEL INSURANCE POLICIES CONFERENCE & EVENT ORGANISERS SCHEDULE SPECIMEN ONLY COMPLETE COVER - `ADD ON' TO TRAVEL INSURANCE POLICIES CONFERENCE & EVENT ORGANISERS SCHEDULE Policy Numer: Policyholder: Certificte of Insurnce: PPIP(E) Period of Insurnce: Finncil Loss:

More information

Grain Marketing: Using Balance Sheets

Grain Marketing: Using Balance Sheets 1 Fct Sheet 485 Grin Mrketing: Using Blnce Sheets Introduction Grin lnce sheets re estimtes of supply nd demnd. They re the key to understnding the grin mrkets. A grin frmer who understnds how to interpret

More information

The phases of a simple compiler: Compiler Construction SMD163. From Intermediate To Target: An Optimizing Compiler: Lecture 13: Instruction selection

The phases of a simple compiler: Compiler Construction SMD163. From Intermediate To Target: An Optimizing Compiler: Lecture 13: Instruction selection Compiler Constrution SMD163 The phses of simple ompiler: Lexer Prser Stti Anlysis IA32 Code Genertor Leture 13: Instrution seletion Viktor Leijon & Peter Jonsson with slides y John Nordlnder Contins mteril

More information

Earning Money. Earning Money. Curriculum Ready ACMNA: 189.

Earning Money. Earning Money. Curriculum Ready ACMNA: 189. Erning Money Curriculum Redy ACMNA: 189 www.mthletics.com Erning EARNING Money MONEY Different jos py different mounts of moneys in different wys. A slry isn t pid once in yer. It is pid in equl prts

More information

Interacting with mathematics in Key Stage 3. Year 9 proportional reasoning: mini-pack

Interacting with mathematics in Key Stage 3. Year 9 proportional reasoning: mini-pack Intercting with mthemtics in Key Stge Yer 9 proportionl resoning: mini-pck Intercting with mthemtics Yer 9 proportionl resoning: mini-pck Crown copyright 00 Contents Yer 9 proportionl resoning: smple unit

More information

Option exercise with temptation

Option exercise with temptation Economic Theory 2008) 34: 473 501 DOI 10.1007/s00199-006-0194-3 RESEARCH ARTICLE Jinjun Mio Option exercise with tempttion Received: 25 Jnury 2006 / Revised: 5 December 2006 / Published online: 10 Jnury

More information

Research Article Existence of Positive Solution to Second-Order Three-Point BVPs on Time Scales

Research Article Existence of Positive Solution to Second-Order Three-Point BVPs on Time Scales Hindwi Publishing Corportion Boundry Vlue Problems Volume 2009, Article ID 685040, 6 pges doi:10.1155/2009/685040 Reserch Article Existence of Positive Solution to Second-Order hree-point BVPs on ime Scles

More information

Read Section 11.1 to Section 11.5 in the textbook. Read instructions on how to install strain gages.

Read Section 11.1 to Section 11.5 in the textbook. Read instructions on how to install strain gages. FABRICATION OF A FORCE TRANSDUCER PREPARATION: Red Section. to Section.5 the textook. Red structions on how to stll str gges. OBJECTIVE: To mke force trnsducer sed upon the endg of cntilever em nd two

More information

Ricardian Model. Mercantilism: 17 th and 18 th Century. Adam Smith s Absolute Income Hypothesis, End of 18 th Century: Major Shift in Paradigm

Ricardian Model. Mercantilism: 17 th and 18 th Century. Adam Smith s Absolute Income Hypothesis, End of 18 th Century: Major Shift in Paradigm Mercntilism: th nd th Century Ricrdin Model lesson in Comprtive dvntge Trde ws considered s Zero-Sum Gme It ws viewed mens to ccumulte Gold & Silver Exports were encourged Imports were discourged End of

More information

CS 188 Introduction to Artificial Intelligence Fall 2018 Note 4

CS 188 Introduction to Artificial Intelligence Fall 2018 Note 4 CS 188 Introduction to Artificil Intelligence Fll 2018 Note 4 These lecture notes re hevily bsed on notes originlly written by Nikhil Shrm. Non-Deterministic Serch Picture runner, coming to the end of

More information

OPEN BUDGET QUESTIONNAIRE UKRAINE

OPEN BUDGET QUESTIONNAIRE UKRAINE Interntionl Budget Prtnership OPEN BUDGET QUESTIONNAIRE UKRAINE September 28, 2007 Interntionl Budget Prtnership Center on Budget nd Policy Priorities 820 First Street, NE Suite 510 Wshington, DC 20002

More information

Managing uncertainty at airports ground handling

Managing uncertainty at airports ground handling AUN014 : Airports in Urn Networks 15-16 Apr 014 CNIT - Pris l Défense (Frnce) Plese Mnging uncertinty t irports ground hndling Author Slm Fitouri-Trelsi, 1 (presenting) Nme, MAIAA/ENAC, First nme*, Toulouse,

More information

Open Space Allocation and Travel Costs

Open Space Allocation and Travel Costs Open Spce Alloction nd Trvel Costs By Kent Kovcs Deprtment of Agriculturl nd Resource Economics University of Cliforni, Dvis kovcs@priml.ucdvis.edu Pper prepred for presenttion t the Americn Agriculturl

More information

AS and A Level Economics A

AS and A Level Economics A AS nd A Level Economics A Getting Strted Guide Person Edexcel Level 3 Advnced GCE in Economics A (9EC0) Person Edexcel Level 3 Advnced Susidiry GCE in Economics A (8EC0) Getting Strted: AS nd A level Economics

More information

3. Argumentation Frameworks

3. Argumentation Frameworks 3. Argumenttion Frmeworks Argumenttion current hot topic in AI. Historiclly more recent thn other pproches discussed here. Bsic ide: to construct cceptble set(s) of beliefs from given KB: 1 construct rguments

More information

OPEN BUDGET QUESTIONNAIRE CZECH REPUBLIC

OPEN BUDGET QUESTIONNAIRE CZECH REPUBLIC Interntionl Budget Project OPEN BUDGET QUESTIONNAIRE CZECH REPUBLIC October 2005 Interntionl Budget Project Center on Budget nd Policy Priorities 820 First Street, NE Suite 510 Wshington, DC 20002 www.interntionlbudget.org

More information

Optimal Revenue-Sharing Double Auctions with Applications to Ad Exchanges

Optimal Revenue-Sharing Double Auctions with Applications to Ad Exchanges Optiml Revenue-Shring Doule Auctions with Applictions to Ad Exchnges Rento Gomes Toulouse School of Economics rento.gomes@tse-fr.eu Vh Mirrokni Google Reserch, ew York mirrokni@google.com ABSTRACT E-commerce

More information

2x2x2 Heckscher-Ohlin-Samuelson (H-O-S) Model with no factor substitution

2x2x2 Heckscher-Ohlin-Samuelson (H-O-S) Model with no factor substitution 2x2x2 Heckscher-Ohlin-Smuelson (H-O-S) odel with no fctor substitution In the H-O-S model we will focus entirely on the role of fctor sulies nd ssume wy differences in technology. Therefore, unlike in

More information

Time Scales: From Nabla Calculus to Delta Calculus and Vice Versa via Duality

Time Scales: From Nabla Calculus to Delta Calculus and Vice Versa via Duality Interntionl Journl of Difference Equtions ISSN 0973-6069, Volume 5, Number 1, pp. 25 40 (2010) http://cmpus.mst.edu/ijde Time Scles: From Nbl Clculus to Delt Clculus nd Vice Vers vi Dulity M. Cristin Cputo

More information

DYNAMIC PROGRAMMING REINFORCEMENT LEARNING. COGS 202 : Week 7 Presentation

DYNAMIC PROGRAMMING REINFORCEMENT LEARNING. COGS 202 : Week 7 Presentation DYNAMIC PROGRAMMING REINFORCEMENT LEARNING COGS 202 : Week 7 Preenttion OUTLINE Recp (Stte Vlue nd Action Vlue function) Computtion in MDP Dynmic Progrmming (DP) Policy Evlution Policy Improvement Policy

More information

ECON 105 Homework 2 KEY Open Economy Macroeconomics Due November 29

ECON 105 Homework 2 KEY Open Economy Macroeconomics Due November 29 Instructions: ECON 105 Homework 2 KEY Open Economy Mcroeconomics Due Novemer 29 The purpose of this ssignment it to integrte the explntions found in chpter 16 ok Kennedy with the D-S model nd the Money

More information

arxiv: v2 [math.nt] 23 Sep 2008

arxiv: v2 [math.nt] 23 Sep 2008 STATISTICAL PROPERTIES OF THE CALKIN WILF TREE: REAL AN p ADIC DISTRIBUTION rxiv:8.54v [mth.nt] 3 Sep 8 GIEDRIUS ALKAUSKAS, JÖRN STEUDING Astrct. We exmine sttisticl properties of the Clkin Wilf tree nd

More information

Class Overview. Database Design. Database Design. Database Design Process. Entity / Relationship Diagrams. Introduction to Database Systems CSE 414

Class Overview. Database Design. Database Design. Database Design Process. Entity / Relationship Diagrams. Introduction to Database Systems CSE 414 Introution to Dtse Systems CSE 44 Leture 9: E/R Digrms Clss Overview Unit : Intro Unit : Reltionl Dt Moels n Query Lnguges Unit : Non-reltionl t Unit 4: RDMBS internls n query optimiztion Unit 5: Prllel

More information

EFFECTS OF THE SINGLE EUROPEAN MARKET ON WELFARE OF THE PARTICIPATING COUNTRIES (theoretical approach)

EFFECTS OF THE SINGLE EUROPEAN MARKET ON WELFARE OF THE PARTICIPATING COUNTRIES (theoretical approach) EFFECTS OF THE SINGLE EUROEAN MARKET ON ELFARE OF THE ARTICIATING COUNTRIES theoretil pproh O ELHD&DU\:DUVDZ6FKRRORIFRRLFV:DUVDZRODG Agnieszk Rusinowsk, rsw Shool of Eonomis, rsw, olnd In this note we

More information

)''/?\Xck_

)''/?\Xck_ bcbsnc.com Deductible options: $250, $500, $1,000 or $2,500 Deductible options $500, $1,000, $2,500, $3,500 or $5,000 D or (100% coinsurnce is not vilble on the $2,500 deductible option) coinsurnce plns:

More information

MATH 236 ELAC MATH DEPARTMENT FALL 2017 SHORT ANSWER. Write the word or phrase that best completes each statement or answers the question.

MATH 236 ELAC MATH DEPARTMENT FALL 2017 SHORT ANSWER. Write the word or phrase that best completes each statement or answers the question. MATH 236 ELAC MATH DEPARTMENT FALL 2017 TEST 1 REVIEW SHORT ANSWER. Write the word or phrse tht best completes ech sttement or nswers the question. 1) The supply nd demnd equtions for certin product re

More information