<?xml version="1.0" standalone="yes"?>
<Paper uid="P04-1074">
  <Title>Applying Machine Learning to Chinese Temporal Relation Resolution</Title>
  <Section position="3" start_page="0" end_page="1" type="intro">
    <SectionTitle>
2 Modeling Temporal Relations
</SectionTitle>
    <Paragraph position="0"/>
    <Section position="1" start_page="0" end_page="0" type="sub_section">
      <SectionTitle>
2.1 Temporal Relation Representations
</SectionTitle>
      <Paragraph position="0"> As the importance of temporal information processing has become apparent, a variety of temporal systems have been introduced, attempting to accommodate the characteristics of relative temporal information. Among those who worked on temporal relation representations, many took the work of Reichenbach (Reichenbach, 1947) as a starting point, while some others based their works on Allen's (Allen, 1981).</Paragraph>
      <Paragraph position="1"> Reichenbach proposed a point-based temporal theory. This was later enhanced by Bruce who defined seven relative temporal relations (Bruce. 1972). Given two durative events, the interval relations between them were modeled by the order between the greatest lower bounding points and least upper bounding points of the two events. In the other camp, instead of adopting time points, Allen took intervals as temporal primitives and introduced thirteen basic binary relations. In this interval-based theory, points are relegated to a subsidiary status as 'meeting places' of intervals. An extension to Allen's theory, which treated both points and intervals as primitives on an equal footing, was later investigated by Ma and Knight (Ma and Knight, 1994).</Paragraph>
      <Paragraph position="2"> In natural language, events can either be punctual (e.g. Bao Zha (explore)) or durative (e.g. Gai Lou (built a house)) in nature. Thus Ma and Knight's model is adopted in our work (see Figure 1). Taking the sentence &amp;quot;Xiu Cheng Li Jiao Qiao Yi Hou ,Ta Men Jie Jue Liao Gai Shi De Jiao Tong Wen Ti (after the street bridge had been built, they solved the traffic problem of the city)&amp;quot; as an example, the relation held between building the bridge (i.e. an interval) and solving the problem (i.e. a point) is</Paragraph>
    </Section>
    <Section position="2" start_page="0" end_page="1" type="sub_section">
      <SectionTitle>
intervals
2.2 Linguistic Features for Determining Relative
Relations
</SectionTitle>
      <Paragraph position="0"> Relative relations are generally determined by tense/aspect, connecting words (temporal or otherwise) and event classes.</Paragraph>
      <Paragraph position="1"> Tense/Aspect in English is manifested by verb inflections. But such morphological variations are inapplicable to Chinese verbs; instead, they are conveyed lexically (Li and Wong, 2002). In other words, tense and aspect in Chinese are expressed using a combination of time words, auxiliaries, temporal position words, adverbs and prepositions, and particular verbs.</Paragraph>
      <Paragraph position="2"> Temporal Connectives in English primarily involve conjunctions, e.g. after, before and when (Dorr and Gaasterland, 2002). They are key components in discourse structures. In Chinese, however, conjunctions, conjunctive adverbs, prepositions and position words are required to represent connectives. A few verbs which express cause and effect also imply a forward movement of event time. The words, which contribute to the tense/aspect and temporal connective expressions, are explicit in a sentence and generally known as Temporal Indicators.</Paragraph>
      <Paragraph position="3"> Event Class is implicit in a sentence. Events can be classified according to their inherent temporal characteristics, such as the degree of telicity and/or atomicity (Li and Wong, 2002). The four widespread accepted temporal classes  are state, process, punctual event and developing event. Based on their classes, events interact with the tense/aspect of verbs to define the temporal relations between two events. Temporal indicators and event classes are together referred to as Linguistic Features (see Table 1). For example, linguistic features are underlined in the sentence &amp;quot;(Yin Wei )Xiu Cheng Li Jiao Qiao (Yi Hou ),Ta Men Jie Jue Liao Gai Shi De Jiao Tong Wen Ti after/because the street bridge had been built (i.e. a developing event), they solved the traffic problem of the city (i.e. a punctual event)&amp;quot;.</Paragraph>
      <Paragraph position="4">  Temporal classification refers to aspectual classification. A punctual event (i.e. represented in time point) A durative event (i.e. represented in time interval)  Table 1 shows the mapping between a temporal indicator and its effects. Notice that the mapping is not one-to-one. For example, adverbs affect tense/aspect as well as discourse structure. For another example, tense/aspect can be affected by auxiliary words, trend verbs, etc. This shows that classification of temporal indicators based on part-of-speech (POS) information alone cannot determine relative temporal relations.</Paragraph>
    </Section>
  </Section>
class="xml-element"></Paper>