<?xml version="1.0" standalone="yes"?>
<Paper uid="W06-2922">
  <Title>Experiments with a Multilanguage Non-Projective Dependency Parser</Title>
  <Section position="5" start_page="166" end_page="166" type="metho">
    <SectionTitle>
3 Inductive Deterministic Parsing
</SectionTitle>
    <Paragraph position="0"> The parser constructs dependency trees employing a deterministic bottom-up algorithm which performs Shift/Reduce actions while analyzing input sentences in left-to-right order.</Paragraph>
    <Paragraph position="1"> Using a notation similar to (Nivre and Scholz, 2003), the state of the parser is represented by a quadruple [?]S, I, T, A[?], where S is the stack, I is the list of (remaining) input tokens, T is a stack of temporary tokens and A is the arc relation for the dependency graph.</Paragraph>
    <Paragraph position="2"> Given an input string W, the parser is initialized to [?](), W, (), ()[?], and terminates when it reaches a configuration [?]S, (), (), A[?].</Paragraph>
    <Paragraph position="3"> The parser by Yamada and Matsumoto (2003) used the following actions: Shift in a configuration [?]S, n|I, T, A[?], pushes n to the stack, producing the configuration [?]n|S, I, T, A[?].</Paragraph>
    <Paragraph position="4"> Right1 in a configuration [?]s1|S, n|I, T, A[?], adds an arc from s1 to n and pops s1 from the stack, producing the configuration [?]S, n|I, T, A[?]{(s1, r, n)}[?].</Paragraph>
    <Paragraph position="5"> Left in a configuration [?]s1|S, n|I, T, A[?], adds an arc from n to s1, pops n from input, pops s1 from the stack and moves it back to I, producing the configuration [?]S, s1|I, T, A[?]{(n, r, s1)}[?].</Paragraph>
    <Paragraph position="6"> At each step the parser uses classifiers trained on treebank data in order to predict which action to perform and which dependency label to assign given the current configuration.</Paragraph>
  </Section>
  <Section position="6" start_page="166" end_page="167" type="metho">
    <SectionTitle>
4 Non-Projective Relations
</SectionTitle>
    <Paragraph position="0"> For handling non-projective relations, Nivre and Nilsson (2005) suggested applying a pre-processing step to a dependency parser, which consists in lifting non-projective arcs to their head repeatedly, until the tree becomes pseudo-projective.</Paragraph>
    <Paragraph position="1"> A post-processing step is then required to restore the arcs to the proper heads.</Paragraph>
    <Paragraph position="2"> 1 Nivre and Scholz reverse the direction, while I follow here the terminology in Yamada and Matsumoto (2003).</Paragraph>
    <Paragraph position="3">  I adopted a novel approach, which consists in adding six new parsing actions: Right2 in a configuration [?]s1|s2|S, n|I, T, A[?], adds an arc from s2 to n and removes s2 from the stack, producing the configuration [?]s1|S, n|I, T, A[?]{(s2, r, n)}[?]. Left2 in a configuration [?]s1|s2|S, n|I, T, A[?], adds an arc from n to s2, pops n from input, pops s1 from the stack and moves it back to I, producing the configuration [?]s2|S, s1|I, T, A[?]{(n, r, s2)}[?].</Paragraph>
    <Paragraph position="4"> Right3 in a configuration [?]s1|s2|s3|S, n|I, T, A[?], adds an arc from s3 to n and removes s3 from the stack, producing the configuration [?]s1|s2|S, n|I, T, A[?]{(s3, r, n)}[?]. Left3 in a configuration [?]s1|s2|s3|S, n|I, T, A[?], adds an arc from n to s3, pops n from input, pops s1 from the stack and moves it back to I, producing the configuration [?]s2|s3|S, s1|I, T, A[?]{(n, r, s3)}[?].</Paragraph>
    <Paragraph position="5"> Extract in a configuration [?]s1|s2|S, n|I, T, A[?], move s2 from the stack to the temporary stack, then Shift, producing the configuration [?]n|s1|S, I, s2|T, A[?].</Paragraph>
    <Paragraph position="6"> Insert in a configuration [?]S, I, s1|T, A[?], pops s1 from T and pushes it to the stack, producing the configuration [?]s1|S, I, T, A[?]. The actions Right2 and Left2 are sufficient to handle almost all cases of non-projectivity: for instance the training data for Czech contain 28081 non-projective relations, of which 26346 can be handled by Left2/Right2, 1683 by Left3/Right3 and just 52 require Extract/Insert. null Here is an example of non-projectivity that can be handled with Right2 (nejen - ale) and Left3 (fax - Vetsinu): Vetsinu techto pristroju lze take pouzivat nejen jako fax, ale soucasne ...</Paragraph>
    <Paragraph position="7"> The remaining cases are handled with the last two actions: Extract is used to postpone the creation of a link, by saving the token in a temporary stack; Insert restores the token from the temporary stack and resumes normal processing.</Paragraph>
    <Paragraph position="8"> This fragment in Dutch is dealt by performing an Extract in configuration [?]moeten|gemaakt|zou, worden|in, A[?] followed immediately by an Insert, leading to the following configuration, which can be handled by normal Shift/Reduce actions: Another linguistic phenomenon is the anticipation of pronouns, like in this Portuguese fragment: Tudo e possivel encontrar em o IX Salao de Antiguidades, desde objectos de ouro e prata, moedas, ...</Paragraph>
    <Paragraph position="9"> The problem here is due to the pronoun Tudo (Anything), which is the object of encontrar (find), but which is also the head of desde (from) and its preceding comma. In order to be able to properly link desde to Tudo, it is necessary to postpone its processing; hence it is saved with Extract to the temporary stack and put back later in front of the comma with Insert. In fact the pair Extract/Insert behaves like a generalized Rightn/Leftn, when n is not known. As in the example, except for the case where n=2, it is difficult to predict the value of n, since there can be an arbitrary long sequence of tokens before reaching the position where the link can be inserted.</Paragraph>
  </Section>
class="xml-element"></Paper>