%Aigaion2 BibTeX export from Idiap Publications
%Sunday 05 May 2024 05:05:05 AM

@INPROCEEDINGS{grangier:2005:nips_workshop,
         author = {Grangier, David and Bengio, Samy},
       projects = {Idiap},
          month = {12},
          title = {Exploiting Hyperlinks to Learn a Retrieval Model},
      booktitle = {NIPS Workshop on Learning to Rank},
           year = {2005},
        address = {Whistler, Canada},
       crossref = {grangier:2005:idiap-05-21},
       abstract = {Information Retrieval (IR) aims at solving a ranking problem: given a query $q$ and a corpus $C$, the documents of $C$ should be ranked such that the documents relevant to $q$ appear above the others. This task is generally performed by ranking the documents $d \in C$ according to their similarity with respect to $q$, $sim (q,d)$. The identification of an effective function $a,b \to sim(a,b)$ could be performed using a large set of queries with their corresponding relevance assessments. However, such data are especially expensive to label, thus, as an alternative, we propose to rely on hyperlink data which convey analogous semantic relationships. We then empirically show that a measure $sim$ inferred from hyperlinked documents can actually outperform the state-of-the-art {\em Okapi} approach, when applied over a non-hyperlinked retrieval corpus.},
            pdf = {https://publications.idiap.ch/attachments/reports/2005/grangier-nips-ranking-workshop.pdf},
     postscript = {ftp://ftp.idiap.ch/pub/reports/2005/grangier-nips-ranking-workshop.ps.gz},
ipdmembership={speech},
}



crossreferenced publications: 
@TECHREPORT{grangier:2005:idiap-05-21,
         author = {Grangier, David and Bengio, Samy},
       projects = {Idiap},
          title = {Inferring Document Similarity from Hyper-links},
           type = {Idiap-RR},
         number = {Idiap-RR-21-2005},
           year = {2005},
    institution = {IDIAP},
       abstract = {Assessing semantic similarity between text documents is a crucial aspect in Information Retrieval systems. In this paper, we propose a technique to derive a similarity measure from hyper-link information. As linked documents are generally semantically closer than unlinked documents, we use a training corpus with hyper-links to infer a function $a,b \to sim(a,b)$ that assigns a higher value to linked documents than to unlinked ones. Two sets of experiments on different corpora show that this function compares favorably with {\em OKAPI} matching on document retrieval tasks.},
            pdf = {https://publications.idiap.ch/attachments/reports/2005/grangier-rr-05-21.pdf},
     postscript = {ftp://ftp.idiap.ch/pub/reports/2005/grangier-rr-05-21.ps.gz},
ipdmembership={speech},
}