Uploaded image for project: 'Solr'
  1. Solr
  2. SOLR-10811

Speed up MultipleAdditiveTreesModel by using QuickScorer algorithm

    XMLWordPrintableJSON

Details

    • Improvement
    • Status: Open
    • Minor
    • Resolution: Unresolved
    • None
    • None
    • contrib - LTR
    • None

    Description

      QuickScorer is an algorithm which can calculate multiple additive trees fast by using bitvectors for detecting target leaves.
      It was first published in SIGIR 2015 and won the best paper award of the conference.
      refs: http://zola.di.unipi.it/rossano/wp-content/papercite-data/pdf/sigir15.pdf

      We implemented QuickScorer as one of LTRScoringModel.
      This model uses same configuration of MultipleAdditiveTreesModel, thus it is easy to replace the model.
      Our experiments show our model can calculate scores about twice faster than MultipleAdditiveTreesModel.

      Attachments

        1. SOLR-10811.patch
          53 kB
          Yuki Yano
        2. SOLR-10811_master.patch
          53 kB
          Yuki Yano
        3. quickscorer_model.pdf
          457 kB
          Yuki Yano

        Issue Links

          is related to

          New Feature - A new feature of the product, which has yet to be developed. SOLR-8542 Integrate Learning to Rank into Solr

          • Major - Major loss of function.
          • Resolved
          relates to

          Task - A task that needs to be done. SOLR-15315 MultipleAdditiveTreesModel Latency optimisation for LTR tree traversal

          • Major - Major loss of function.
          • Closed

          Activity

            People

              Unassigned Unassigned
              yuyano Yuki Yano
              Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

              Dates

                Created:
                Updated: