<?xml version="1.0" standalone="yes"?>
<Paper uid="P06-2087">
  <Title>Argumentative Feedback: A Linguistically-motivated Term Expansion for Information Retrieval</Title>
  <Section position="4" start_page="675" end_page="676" type="intro">
    <SectionTitle>
2 Related works
</SectionTitle>
    <Paragraph position="0"> Our basic experimental hypothesis is that some particular sentences, selected based on argumentative categories, can be more useful than others to support well-known feedback information retrieval tasks. It means that selecting sentences based on argumentative categories can help focusing on content-bearing sections of scientific articles.</Paragraph>
    <Section position="1" start_page="675" end_page="675" type="sub_section">
      <SectionTitle>
2.1 Argumentation
</SectionTitle>
      <Paragraph position="0"> Originally inspired by corpus linguistics studies (Orasan, 2001), which suggests that scientific reports (in chemistry, linguistics, computer sciences, medicine...) exhibit a very regular logical distribution -confirmed by studies conducted on biomedical corpora (Swales, 1990) and by ANSI/ISO professional standards - the argumentative model we experiment is based on four disjunct classes: PURPOSE, METHODS, RE-SULTS, CONCLUSION.</Paragraph>
      <Paragraph position="1"> Argumentation belongs to discourse analysis1, with fairly complex computational models such as the implementation of the rhetorical structure theory proposed by (Marcu, 1997), which proposes dozens of rhetorical classes.</Paragraph>
      <Paragraph position="2"> More recent advances were applied to document summarization. Of particular interest for our approach, Teufel and Moens (Teufel and Moens, 1999) propose using a list of manually crafted triggers (using both words and expressions such as we argued, in this article, the paper is an attempt to, we aim at, etc.) to automatically structure scientific articles into a lighter model, with only seven categories: BACKGROUND, TOPIC, RELATED WORK, PURPOSE, METHOD, RESULT, and CONCLUSION. null More recently and for knowledge discovery in molecular biology, more elaborated models were proposed by (Mizuta and Collier, 2004) (Mizuta et al., 2005) and by (Lisacek et al., 2005) for novelty-detection. (McKnight and Srinivasan, 2003) propose a model very similar to our four-class model but is inspired by clinical trials. Preliminary applications were proposed for bib- null appropriate argumentative distribution belong to logics, while ill-defined ones belong to rhetorics.</Paragraph>
      <Paragraph position="3"> liometrics and related-article search (Tbahriti et al., 2004) (Tbahriti et al., 2005), information extraction and passage retrieval (Ruch et al., 2005b). In these studies, sentences were selected as the basic classification unit in order to avoid as far as possible co-reference issues (Hirst, 1981), which hinder readibity of automatically generated and extracted sentences.</Paragraph>
    </Section>
    <Section position="2" start_page="675" end_page="676" type="sub_section">
      <SectionTitle>
2.2 Query expansion
</SectionTitle>
      <Paragraph position="0"> Various query expansion techniques have been suggested to provide a better match between user information needs and documents, and to increase retrieval effectiveness. The general principle is to expand the query using words or phrases having a similar or related meaning to those appearing in the original request. Various empirical studies based on different IR models or collections have shown that this type of search strategy should usually be effective in enhancing retrieval performance. Scheme propositions such as this should consider the various relationships between words as well as term selection mechanisms and term weighting schemes (Robertson, 1990). The specific answers found to these questions may vary; thus a variety of query expansion approaches were suggested (Efthimiadis, 1996).</Paragraph>
      <Paragraph position="1"> In a first attempt to find related search terms, we might ask the user to select additional terms to be included in a new query, e.g. (Velez et al., 1997). This could be handled interactively through displaying a ranked list of retrieved items returned by the first query. Voorhees (Voorhees, 1994) proposed basing a scheme based on the WordNet thesaurus. The author demonstrated that terms having a lexical-semantic relation with the original query words (extracted from a synonym relationship) provided very little improvement (around 1% when compared to the original unexpanded query).</Paragraph>
      <Paragraph position="2"> As a second strategy for expanding the original query, Rocchio (Rocchio, 1971) proposed accounting for the relevance or irrelevance of top-ranked documents, according to the user's manual input. In this case, a new query was automatically built in the form of a linear combination of the term included in the previous query and terms automatically extracted from both the relevant documents (with a positive weight) and non-relevant items (with a negative weight). Empirical studies (e.g., (Salton and Buckley, 1990)) demonstrated that such an approach is usually quite effective, and could  be used more than once per query (Aalbersberg, 1992). Buckley et al. (Singhal et al., 1996b) suggested that we could assume, withoutevenlookingatthemoraskingtheuser, that the top k ranked documents are relevant. Denoted the pseudo-relevance feedback or blindquery expansion approach, this approach is usually effective, at least when handling relatively large text collections.</Paragraph>
      <Paragraph position="3"> As a third source, we might use large text corpora to derive various term-term relationships, using statistically or information-based measures (Jones, 1971), (Manning and Sch&amp;quot;utze, 2000). For example, (Qiu and Frei, 1993) suggested that terms to be added to a new query could be extracted from a similarity thesaurus automatically built through calculating co-occurrence frequencies in the search collection. The underlying effect was to add idiosyncratic terms to the underlying document collection, related to the query terms by language use. When using such query expansion approaches, we can assume that the new terms are more appropriate for the retrieval of pertinent items than are lexically or semantically related terms provided by a general thesaurus or dictionary. To complement this global document analysis, (Croft, 1998) suggested that text passages (with a text window size of between 100 to 300 words) be taken into account. This local document analysis seemed to be more effective than a global term relationship generation.</Paragraph>
      <Paragraph position="4"> As a forth source of additional terms, we might account for specific user information needs and/or the underlying domain. In this vein, (Liu and Chu, 2005) suggested that terms related to the user's intention or scenario might be included. In the medical domain, it was observed that users looking for information usually have an underlying scenario in mind (or a typical medical task). Knowing that the number of scenarios for a user is rather limited (e.g., diagnosis, treatment, etiology), the authors suggested automatically building a semantic network based on a domain-specific thesaurus (using the Unified Medical Language System (UMLS) in this case). The effectiveness of this strategy would of course depend on the quality and completeness of domain-specific knowledge sources. Using the well-known term frequency (tf)/inverse document frequency (idf) retrieval model, the domain-specific query-expansion scheme suggested by Liu and Chu (2005) produces better retrieval performance than a scheme based on statistics (MAP: 0.408 without query expansion, 0.433 using statistical methods and 0.452 with domain-specific approaches).</Paragraph>
      <Paragraph position="5"> In these different query expansion approaches, various underlying parameters must be specified, and generally there is no single theory able to help us find the most appropriate values. Recent empirical studies conducted in the context of the TREC Genomics track, using the OHSUGEN collection (Hersh, 2005), show that neither blind expansion (Rocchio), nor domain-specific query expansion (thesaurus-based Gene and Protein expansion) seem appropriate to improve retrieval effectiveness (Aronson et al., 2006) (Abdou et al., 2006).</Paragraph>
    </Section>
  </Section>
class="xml-element"></Paper>