<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-0832">
  <Title>SENSEVAL Automatic Labeling of Semantic Roles using Maximum Entropy Models</Title>
  <Section position="3" start_page="1" end_page="5" type="intro">
    <SectionTitle>
2 Model
</SectionTitle>
    <Paragraph position="0"> We separate the problem of FrameNet tagging into three subsequent processes: 1) sentence  segmentation 2) frame element identification, and 3) semantic role tagging. We assume the frame element (FE) boundaries match the parse constituents, so we segment a sentence based on  parse constituents. We consider steps 2) and 3) as classification problems. In frame element identification, we use a binary classifier to determine if each parse constituent is a FE or not, while, in semantic role tagging, we classify each  http://www.icsi.berkeley.edu/~framenet identified FE into its appropriate semantic role.  Figure 1 shows the sequence of steps.</Paragraph>
    <Paragraph position="1"> He fastened the panel from an old radio to the headboard with sticky tape and tied the driving wheel to Pete 's cardboard box with string (He) (fastened the panel from an old radio to the headboard with sticky tape) (and) (tied) (the driving wheel) (to Pete 's  having a target word &amp;quot;tied&amp;quot;.</Paragraph>
    <Paragraph position="2"> We train the ME models using the GIS algorithm (Darroch and Ratcliff, 1972) as implemented in the YASMET ME package (Och, 2002). We use the YASMET MEtagger (Bender et al. 2003) to perform the Viterbi search for choosing the most probable tag sequence for a sentence using the probabilities computed during training. Feature weights are smoothed using Gaussian priors with mean 0 (Chen and Rosenfeld, 1999).</Paragraph>
    <Section position="1" start_page="2" end_page="2" type="sub_section">
      <SectionTitle>
2.1 Sentence Segmentation
</SectionTitle>
      <Paragraph position="0"> We segment a sentence into a sequence of non-overlapping constituents instead of all individual constituents. There are a number of advantages to applying sentence segmentation before FE  We are currently ignoring null instantiations.</Paragraph>
    </Section>
    <Section position="2" start_page="2" end_page="4" type="sub_section">
      <SectionTitle>
Association for Computational Linguistics
</SectionTitle>
      <Paragraph position="0"> for the Semantic Analysis of Text, Barcelona, Spain, July 2004 SENSEVAL-3: Third International Workshop on the Evaluation of Systems boundary identification. First, it allows us to utilize sentence-wide features for FE identification.</Paragraph>
      <Paragraph position="1"> The sentence-wide features, containing dependent information between frame element such as the previously identified class or the syntactic pattern, have previously been shown to be powerful features for role classification (Fleischman et al., 2003). Further, it allows us to reduce the number of candidate constituents for FE, which reduces the convergence time in training.</Paragraph>
      <Paragraph position="2"> The constituents are derived from a syntactic parse tree  . Although we need to consider all combinations of various level constituents in a parse tree, we know the given target word should be a separate segment because a target word is not a part of other FEs.</Paragraph>
      <Paragraph position="3">  Since most frame elements tend to be in higher levels of the parse tree, we decide to use the highest constituents (the parse constituents having the maximum number of words) while separating the target word. Figure 2 shows an example of the segmentation for an actual sentence in FrameNet with the target word &amp;quot;tied&amp;quot;.</Paragraph>
      <Paragraph position="4"> He tied the to Pete box  the target predicate, and the shaded constituent represents each segment.</Paragraph>
      <Paragraph position="5"> However, this segmentation reduces the FE coverage of constituents (the number of constituents matching frame elements). In Table 1, &amp;quot;individual constituents&amp;quot; means a list of all constituents, and &amp;quot;Sentence segmentation&amp;quot; means a sequence of non-overlapping constituents that are taken in our work. We can regard 85.8% as the accuracy of the parser.</Paragraph>
    </Section>
    <Section position="3" start_page="4" end_page="5" type="sub_section">
      <SectionTitle>
2.2 Frame Element Identification
</SectionTitle>
      <Paragraph position="0"> Frame element identification is executed for segments to classify into the classes on FE, Target, or None. When a constituent is both a target and a frame element, we set it as a frame element when training because we are interested in identifying frame elements not a target.</Paragraph>
      <Paragraph position="1"> The initial features are adopted from (Gildea and Juraksky 2002) and (Fleischman, Kwon, and Hovy 2003), and a few additional features are also used.</Paragraph>
      <Paragraph position="2"> The features are: * Target predicate (target): The target is the principal lexical item in a sentence.</Paragraph>
      <Paragraph position="3"> * Target lexical name (lexunit): The formal lexical name of target predicate is the string of the original form of target word and grammatical type. For example, when the target is &amp;quot;tied&amp;quot;, the lexical name is &amp;quot;tie.v&amp;quot;. * Target type (ltype): The target type is a part of lexunit representing verb, noun, or adjective. (e.g. &amp;quot;v&amp;quot; for a lexunit &amp;quot;tie.v&amp;quot;) * Frame name (frame): The semantic frame is defined in FrameNet with corresponding target.</Paragraph>
      <Paragraph position="4"> * Constituent path (path): From the syntactic parse tree of a sentence, we extract the path from each constituent to the target predicate.</Paragraph>
      <Paragraph position="5"> The path is represented by the nodes through which one passes while traveling up the tree from the constituent and then down through the governing category to the target word. For example, &amp;quot;the driving wheel&amp;quot; in the sentence of Figure 2 has the path, NP|VP|VBD.</Paragraph>
      <Paragraph position="6"> * Partial path (ppath): The partial path is a variation of path, and it produces the same path as above if the constituent is under the same &amp;quot;S&amp;quot; as target word, if not, it gives &amp;quot;nopath&amp;quot;. * Syntactic Head (head): The syntactic head of each constituent is obtained based on Michael Collins's heuristic method  . When the head is a proper noun, &amp;quot;proper-noun&amp;quot; substitutes for the real head. The decision as to whether the head is a proper noun is made based on the part of speech tags used in the parse tree.</Paragraph>
      <Paragraph position="7">  (e.g., NP, PP) of each constituent is also extracted from the parse tree. It is not the same as the manually defined PT in FrameNet.</Paragraph>
      <Paragraph position="8"> * Logical Function (lf): The logical functions of constituents in a sentence are simplified into three values: external argument, object argument, other. When the constituent's phrase type is NP, we follow the links in the parse tree from the constituent to the ancestors until we meet either S or VP. If the S is found first, we assign external argument to the constituent, and if the VP is found, we assign object argument. Otherwise, other is assigned.</Paragraph>
      <Paragraph position="9"> * Position (pos): The position indicates whether a constituent appears before or after the target predicate.</Paragraph>
      <Paragraph position="10"> * Voice (voice): The voice of a sentence (active, passive) is determined by a simple regular expression over the surface form of the sentence.</Paragraph>
      <Paragraph position="11"> * Previous class (c_n): The class information of the n th -previous constituent (Target, FE, or None) is used to exploit the dependency between constituents. During training, this information is provided by simply looking at the true class of the constituent occurring npositions before the target element. During testing, the hypothesized classes are used for Viterbi search.</Paragraph>
    </Section>
    <Section position="4" start_page="5" end_page="5" type="sub_section">
      <SectionTitle>
Feature Set Example Functions
</SectionTitle>
      <Paragraph position="0"> f(c, lexunit) f(c, tie.v) = 1 f(c, pt, pos, voice) f(c, NP,after,active) = 1 f(c, pt, lf) f(c, ADVP,obj) = 1 f(c, pt_-1, lf_-1) f(c, VBD_-1, other_-1) = 1 f(c, pt_1, lf_1) f(c, PP_1, other_1) = 1 f(c, head) f(c, wheel) = 1 f(c, head, frame) f(c, wheel, Attaching) = 1 f(c, path) f(c, NP|VP|VBD) = 1 f(c, path_-1) f(c, VBD_-1) = 1 f(c, path_1) f(c, PP|VP|VBD_1) = 1 f(c, target) f(c, tied) = 1 f(c, ppath) f(c, NP|VP|VBD) = 1 f(c, ppath, pos) f(c,NP|VP|VBD, after) = 1 f(c, ppath_-1, pos_-1) f(c, VBD_-1,after) = 1 f(c ,ltype, ppath) f(c, v, NP|VP|VBD) = 1 f(c ,ltype, path) f(c, v, NP|VP|VBD) = 1 f(c ,ltype, path_-1) f(c, v,VBD_-1) = 1 f(c frame) f(c, Attaching) = 1 f(c, frame, c_-1) f(c, Attaching, T_-1) = 1 f(c,frame, c_-2,c_-1) f(c, Attaching,NO_-2,T_-1)=1 Table 2. Feature sets used in ME frame element identification. Example functions of &amp;quot;the driving wheel&amp;quot; from the sample sentence in Fig.2. The combinations of these features that are used in the ME model are shown in Table 2. These feature sets contain the previous or next constituent's features, for example, pt_-1 represents the previous constituent's phrase type and lf_1 represents the next constituent's logical function.</Paragraph>
    </Section>
    <Section position="5" start_page="5" end_page="5" type="sub_section">
      <SectionTitle>
2.3 Semantic Role Classification
</SectionTitle>
      <Paragraph position="0"> Semantic role classification is executed only for the constituents that are classified into FEs in the previous FE identification phase by employing Maximum Entropy classification.</Paragraph>
      <Paragraph position="1"> In addition to the features in Section 2.2, two more features are applied.</Paragraph>
      <Paragraph position="2"> * Order (order): The relative position of a frame element in a sentence is given. For example, the sentence from Figure 2 has four frame elements, where the element &amp;quot;He&amp;quot; has order 0, while &amp;quot;with string&amp;quot; has order 3. * Syntactic pattern (pat): The sentence level syntactic pattern is generated from the parse tree by considering the phrase type and logical functions of each frame element in the sentence. In the example sentence in Figure 2, &amp;quot;He&amp;quot; is an external argument Noun Phrase, &amp;quot;tied&amp;quot; is a target predicate, and &amp;quot;the driving wheel&amp;quot; is an object argument Noun Phrase.</Paragraph>
      <Paragraph position="3"> Thus, the syntactic pattern associated with the sentence is [NP-ext, target, NP-obj, PP-other, PP-other].</Paragraph>
      <Paragraph position="4"> Table 3 shows the list of feature sets used for the ME role classification.</Paragraph>
      <Paragraph position="5"> Feature Set f(r, lexunit) f(r, pt, lf) f(r, target) f(r, pt_-1, lf_-1) f(r, pt, pos, voice) f(r, pt_1, lf_1) f(r, head) f(r, order, syn) f(r, head, lexunit) f(r, lexunit, order, syn) f(r, head, frame) f(r, pt, pos, voice, lexunit) f(r, frame, r_-1) f(r, frame, r_-2,r_-1) f(r, frame,r_-3, r_-2,r_-1)</Paragraph>
    </Section>
  </Section>
class="xml-element"></Paper>