File Information
File: 05-lr/acl_arc_1_sum/cleansed_text/xml_by_section/metho/00/c00-1043_metho.xml
Size: 16,363 bytes
Last Modified: 2025-10-06 14:07:09
<?xml version="1.0" standalone="yes"?> <Paper uid="C00-1043"> <Title>References</Title> <Section position="4" start_page="293" end_page="296" type="metho"> <SectionTitle> 3 Interactions </SectionTitle> <Paragraph position="0"> Three main interactions between text-surface-based and knowledge-based NLP techniques are designed in our Q/A architecture: 1. When multiple sets of question keywords are passed to the search engine, increasing the chance of finding the text paragraph containing the answer.</Paragraph> <Paragraph position="1"> 2. When the question focus and the answer type, resuiting from the knowledge-based processing of the question, are used in the extraction of the answer, based on several empirical scores.</Paragraph> <Paragraph position="2"> 3. When the justification option of the Q/A system is available. Instead of returning answers scored by some empirical measures, a proof of the correctness of the answer is produced, by accessing the logical transformations of the question and the answer, as well as axioms encoding world knowledge.</Paragraph> <Paragraph position="3"> All these interactions depend on two Nctors: (1) the l;ransformations of the question or answer into semantic or logical representations; and (2) the availability of knowledge resources, e.g. the question taxonomy and the world knowledge axioms. The availability of new, high-performace parsers that operate on real world texts determines the transformation into semantic and logic formulae quite simple. In addition, the acquisition of question taxonomies is alleviated by machine learning techlfiques inspired from bootstrapping methods that learn linguistic patterns and semantic dictionaries for IE (of. (Riloff and Jones, 1999)). World knowledge axioms can also be easily derived by processing the gloss (lefinitions of WordNel; (Fellbaunl 1998).</Paragraph> <Section position="1" start_page="293" end_page="294" type="sub_section"> <SectionTitle> a.1 Semantic and Logic Transformations Semantic Transtbrmations </SectionTitle> <Paragraph position="0"> Instead of t)roducing only a phrasal parse for the question and answer, we lnake use of one of the new statistical parsers of large real-world text coverage (Collins, 1996). The parse trees produced by such a parser can be easily translated into a seinantic representation that (1) comprises all the phrase beads and (2) captures their int,er-relationships by anonymous links. Figure 2 illustrates both the I)arse tree and the associated semantic representation of a TI{EC-8 question.</Paragraph> <Paragraph position="1"> Question: Why did I)avid Koresh ask the FBI for a word processor?</Paragraph> <Paragraph position="3"> The actual transformation into semantic representation of a question or an answer is obtained as a by-product of the parse tree traversal. Initially, all leaves of the parse tree are classified as .@@nodes or no'n-skipnodes. All n(mns, non-auxiliary verbs, adjectives a.nd adverl)s are categorized as non-skitmodes. All the other h~aves are skipnodes. Bottom-u 1) trav('.rsal of tim 1)arse tree (:ntails tlm t)roi)agation of leaf labels wh('amver l;hc 1)arcnt nod(; has more than one non-skipnod(; child. A rule based on the syntactic category (it' th(.' father selects one of the childr(m to 1)ropagatc its label a,t the next level in the tree. The winning node will then be considered linked to all the other fornmr siblings thai; al'e non-skilmodes. The prot)agation (:ontimms mltil the l)arse 1;l'(~.c root receives a label, and thus a scmanti(&quot; gral)h is (:rc;tl;(;(l as a 1)y-1)rodu(:t. Part of th('. label i)roI)agation, we also (:onsider that whenever all ('hildr(;n of a non-terminal are skilmo(l(;,% the parent; becomes a. skipnode as well.</Paragraph> <Paragraph position="4"> Figure 3 rel)r(~'s(;nts the lal)el I)rOl)agation for th(; 1)arse tree of the question l'el)res(mt(M in Figure 2.</Paragraph> <Paragraph position="5"> The labels of Korcsh, ask, FB1 and processor are l)rot)agated to tlw, next level. This entails t;hat Kor'c.s'h is linked to David, ask to I,'BI and procc',s'sor mM procc, ssor to word. As a.@ be(:(/m(!s the lab(:l of th(; tree to(it, it is a.lso linked to I~ds'A,9ON, the qu(~sti()n type (l('~I;(n'lnin(',d l)y tlm question st('m: wh, y. The lal)el 1)rot)agation rules are id(mtical t<) the rules fl)r mapping from trees to d(~t)(m(hm(:y s\[.rlt(:l;lll.'cs llSC,(l my Mi('ha(,q Collins (of. (Collins, 199(5)). These rules i(lenti\[}, the head-child, and pl'Ol)agatc its label up in the tree.</Paragraph> <Paragraph position="7"> The logical formulae in whi(:h questions or answers arc translated are inspired \[)y the notal;i(m prol/os('.d in (ltobbs, 1.986-1.) and implemented in TACITUS (ttobbs, 1986-2).</Paragraph> <Paragraph position="8"> Based on the davidsonian tl&quot;eatmenl; of action sen|;(?llC(;S, in which events are tr(~at, ed as individuals, every question and every answer are transform('xl in a tirst-order t)redicatc tbrmula for which (1) verbs are mapped in 1)red\[cares vcvb(e,x,y,z,...) with the.</Paragraph> <Paragraph position="9"> (:onvention thai; variable e rel)res(;nts the evc'nl;uali(;y of that acl;ion or even(; (;o take place, wh('a'eas (;lm othcu arguments (e.g. z, y, z, ...) repr(~,s(mt l;lm t)rcdicate argmnents of the verb; (2) nouns arc mal)l)ed into their lexicalized predicates; raM, (3) mo(lificws have the same argument as the predicate they modi\[y. For (;Xalnpl('~, l;he qu(~si;ion illustrated in Figure 2 has l;he following logical fornl transforlnation (LFT) :</Paragraph> <Paragraph position="11"> The process of trmlslat;ing a sema.ntic ret)resenta.tion into a logic form has the following steps: 1. For each node in the semantic rcprcscntation, create a prcdicatc with a distinct argument.</Paragraph> <Paragraph position="12"> 2.a. If it noun and (tit ad, jective predicate arc linlccd tht:y should have, the sam(&quot; argument.</Paragraph> <Paragraph position="13"> 2.b. Tltc .qamc fin&quot; verbs and adverbs, pairs of lwuns or an adjective and an advcrb.</Paragraph> <Paragraph position="14"> 3. l,b'r each verb predicate, add aT~lumcnts corrc.sponding to each predicate to which it is directly linked in the semantic representation.</Paragraph> <Paragraph position="15"> Predicate argunmnts (:an be identified because tim soma,hi;it l'ol)res(.~nl;al;ion using &nollymous relations represenl;s uniformly adjuncts mM thematic roles. Ilowevel:, sl;e 1) 2 of the l;l&quot;mtsl~d;ion procedure l'eCOgnixes the adjuncts, making predicate argmnenl;s the remaining (:()nn(~(:tions of tlm v(n'}) in t;}l(, ~ semal~I;ic rq)resentation.</Paragraph> </Section> <Section position="2" start_page="294" end_page="296" type="sub_section"> <SectionTitle> 3.2 Question Taxonomy </SectionTitle> <Paragraph position="0"> The question taxonomy rel)rcscnl;s each question nod(', as a quintuple: (1) a setnan(;ic rcpr(~s(!ni;ation of a qucsLion; (2) th(; question type; (3) the m~swer typ(,; (4) the question focus an,l (5) the question keywoMs. By using over 1500 questions provided by \]lcme(lia, as well as otlmr 2000 quest;ions retrieved from FAQFinder, we have b(!en abh ~. to learn classitication rules mM buihl a (:oinplex question taxonomy.</Paragraph> <Paragraph position="2"> Initially, we stm'tcd with a seed hit;rarchy of 25 question classes, manually built, in which all the semantic classes of the nodes fl'om the semantic representations were decided oil-line, by a hmnan expert.</Paragraph> <Paragraph position="3"> 300 questions were processed to create this seed hierarchy. Figure 4 illustrates some of the nodes of the top of this hierarchy. Later, as 500 more questions were considered, we started classifying them semi-automatically, using the following two steps: (1) first a hmnan wonld decide the semantic class of each node in the semantic representation of the new question; (2) then a classification procedure would decide whether the question belongs to one of the existing classes or a new class should be considered.</Paragraph> <Paragraph position="4"> To be able to classify a new question in one of the existing question classes, two conditions must be satisfled: (a) all nodes fi'om the taxonomy question must correspond to new question nodes with the same semantic classes; and (b) unifyable nodes must be linked in the same way in both representations. The hierarchy grew to 68 question nodes.</Paragraph> <Paragraph position="5"> Later, 2700 more questions were classified fully automatically. To decide the semantic classes of the nodes, we used the WordNet semantic hierarchies, by simply assigning to each semantic representation node the same class as that of any other question term from its WordNet hiera.rchy.</Paragraph> <Paragraph position="6"> The semantic representation, having the same format for questions and answers, is a case fi'ame with anonymous relations, that allows the unification of the answer to the question regardless of the case relation. Figure 5 illustrates tbur nodes fi'om the question taxonomy, two for the &quot;currency&quot; question tyI)e attd two for the &quot;person name&quot; question type. The Figure also represents the mappings of four TREC-8 questions in these hier~rchy nodes. The mappings are represented by dashed arcs. In this Figm'c, the nodes front the semantic representations that conrain a question mark are place holders for the expected answer type.</Paragraph> <Paragraph position="7"> An additional set of classification rules is assoeiated with this taxonomy, hfitially, all rules are based on the recognition of the question stem and of the answer type, obtained with class intbrmation from WordNet. However we could learn new rules when inorphologieal and semantic variations of the semantic nodes arc allowed. Moreover, along with the new rules, we enrich the taxonomy, because often the new questions unify only partially with the current taxenemy. All semantic and morphologic variations of the semantic representation nodes are grouped into word classes. Several of the word classes we used are listed in Table 1.</Paragraph> <Paragraph position="8"> \[I Word Ulass Words l\] Value words &quot;monetary value&quot;, &quot;money&quot;, &quot;price&quot; Expenditure words &quot;spend&quot;, &quot;buy&quot;, &quot;rent&quot;, &quot;invest&quot; Creation words &quot;author&quot;, &quot;designer&quot;, ainvent~... The bootstrapping algorithm that learns new classification rules and new classes of questions is based on an intbrmation extraction measure: scorc(rulei)--Ai * lo.q2(Ni), where Ai stands for the number of different lexicon entries for the answer type of the question, whereas Ni = Ai/Fi, where Fi is the number of different focus categories classified. The steps of the bootstrapping algorithm are: 1. Retrieve concepts morphologically//semantically related to the semantic representations 2. Apply the classification rules to all questions that contain any newly retrieved concepts.</Paragraph> <Paragraph position="9"> 3. New_Classificatiton_Rules={} MUTUAL BOOTSTRAPPING LOOP 4. Score. all new classification rules 5. best_CR=thc highest scoring classification rule 6. Add bcst_CR to the classification rules 7. Add the questions classified by best_CR to the taxonomy 8. Goto step 4 three times.</Paragraph> <Paragraph position="10"> 9. Discard all new rules but the best scoring three.</Paragraph> <Paragraph position="11"> 10. Goto 3. until the Q/A performance improves.</Paragraph> </Section> </Section> <Section position="5" start_page="296" end_page="297" type="metho"> <SectionTitle> 4 The Justification Option </SectionTitle> <Paragraph position="0"> A Q/A system that provides with the option of justil~ving the answer has the advantage that erroneous answers can be ruled out syst(,'matieally. In our quest of enh~mcing the precision of a Q/A system by incof porating additional knowledge, we fount1 this option very helpflH. However~ the generation of justifications for ol)en-domain textual Q/A systems poses some challenges. First;, we needed to develol) a very efficient prover, operating on logical form transfermat;ions. Our 1)rool'q are backchaining Do\]n the qllestions through a mixture of axioms. We use thl'ee forlllS of axioms: (1) axioms derived fl'om the facts stated in I;he l;eXtllal allswel; (2) ~/XiOlllS ro4)resenting world knowledge,; and (3) axioms (let;ermined by coreference resohlt;ion in the atiswer text;. For exalni)le , some of the axioms eml)loyed 1;o prove the answer to the TREC~8 question (26: Why did David Kor'csh, ask th, c 1,7\]1 for a 'wo~'d p~vccssor?&quot; are: so'r to c',,ablc h,i'H~, to 'record his vt:'~;clatio'~,s&quot;. The sec-Ol~d set ret)r(',senl;s worl(1 knowledge axioms that; we acquired senfi-auto\]na.tically. For instance we know that David is a male name, thus tlmt person cmt be addressed with Mr.. Similarly, events are (real>led for some reason. The third set of axioms represent the fact that the t,'l~l is in the context of the text answer. To be noted that the axioms derived from the answer have construct argument, s, relnesentc, d by convention with numl)ers larger titan 70. All the other arguments are variables.</Paragraph> <Paragraph position="1"> Q52 \Y=ho invented the road trallic cone,? Answer ST~iliT~g proudly for the caTl~cras , Gover~tor (shallow Pete Wilson, US TraTtsportatioTt Secretary methods) l,'edcrico l)eT~a aTtd Mayor l~ichavd l~iovda~t removed ~t half- dozeTt phtstic or~t~tgc coTtcs firm the road~oay aTtd the first ca~'s passed Tim justification of this answer is t)rovided by the r~ ( following proof I;r~lee. \] h ', 1)rover ;tttelllI)tS t;o 1)rove the LFT of the question (QLF) corre(:t 1} 3, proving from left to right each term of QLF.</Paragraph> <Paragraph position="2"> ->Answer:0ver the weekend Mr Koresh sent a request for a word processor to enable him to record his revelations.</Paragraph> <Paragraph position="4"> There are i target axioms. Selected axiom: David(1):= Mr(1).</Paragraph> <Paragraph position="5"> Unifying: i to i. Ok</Paragraph> <Paragraph position="7"> There are i target axioms. Selected axiom: processor(72):= null.</Paragraph> <Paragraph position="8"> Unifying: ?2 to 72. Ok --> Proving: FBl(4)'ask(3 4 72 71 5)^REASON(5)^_PER(71)'_0RG(4) There are i target axioms, Selected axiom: FBI(1):= null. Unifying: 4 to i. Ok --> Proving: ask(3 4 72 71 5)~REASON(5)^_PER(ZI)^_BRG(4) There are 2 target axioms. Selected axiom: ask(l 2 3 4 5):= sent(l 6 7 4)'request(5).</Paragraph> <Paragraph position="9"> Unifying: i to 2. 3 to i. 5 to 5. 71 to 4. 72 to 3. Ok --> Proving: sent(l 6 7 ?I)^request(6)^_REASON(5)^_PER(YI)^_0RG(2) There are I target axioms, Selected axiom: sent(Z7 76 78 71):= null. Unifying: I to 77. 6 to Y6. Y to 78. 71 to 71. Ok --> Proving: request(76)^REASON(5)'_PER(71)^_0RG(2) There are i target axioms. Selected axiom: request(76):= null. Unifying: 76 to 76. Ok --> Proving: _REASON(5)degPER(71)deg8}~G(2) There are 1 target axioms. Selected axiom: _REASON(5):= enable(5 3 6). Unifying: 5 to 5. flk --> Proving: enable(5 3 6)^_PER(YI)-_flRG(2) There are i target axioms. Selected axiom: enable(75 73 76): = null. Unifying: 3 to ?3. 5 to 75. 6 to 76. Dk --> Proving: _PER(71)^_ORG(2) There are 3 target axioms. Selected axiom: _PER(71):= null. Unifying: 71 to gl. 0k --> Proving: _ORG(2) There are i target axioms. Selected axiom: _0RG(1):= FBI(1). Unifying: 2 to i. Ok --> Proving: nullJl\]J We found:Success.</Paragraph> <Paragraph position="10"> .................................................................. There are cases when our simple prover fails to prove a. correct answer. We have notice(1 that this hal)pens 1)ecause in the answer semantic representation, st)me concepts that are connected in the question semantic representation are no longer directly linked. This is due to the f~mt that there are either parser errors or there are new syntactic dependencies between the two concepts. To acconmmdm;e this situation, we allow diflhrent constants that are arguments of the sanle predicate to be unifiable. The special cases in which this relaxation of the unification i)roeedul'e is allowed constitute our abduction rltles.</Paragraph> </Section> class="xml-element"></Paper>