学会員メーリングリストアーカイブ (2012年)

NTCIR-10参加募集 NTCIR-Mathタ スク


日本データベース学会の皆様

国立情報学研究所の相澤と申します.

情報検索(サーチ)と情報アクセス技術の評価ワークショップ
NTCIR-10における数式検索パイロットタスクのお知らせです.
科学技術論文中の数式検索や,数式定義のアノテーション
データを使ったタスクを企画しています.ご興味のある方は
どうぞ参加をご検討下さい.

##########################################################
          CALL FOR TASK PARTICIPATION
            NTCIR10-Tasks: Math Task:
Math Retrieval Subtask and Math Understanding Subtask
Http://ntcir-math.nii.ac.jp/

Topics released: October 2012
Evaluation results released: February 2013
Final NTCIR-10 workshop: June 18-21 2013
##########################################################


### OVERVIEW ###

Mathematical formulae are important means for dissemination and
communication of scientific information. They are not only used for
numerical calculation but also for clarifying definitions or
disambiguating explanations that are written in natural language. The
NTCIR Math Task aims to explore search methods tailored to mathematical
content through the design of suitable search tasks and the construction
of evaluation datasets.

Despite the importance of mathematical modeling and analysis in many
scientific disciplines, most of the contemporary retrieval systems do
not facilitate users' access to mathematical expressions in target
scientific documents. Researchers that are eager to work on this problem
are faced with two major obstacles. First, there exists only a few
readily available resources with mathematical structured information.
Second, there are neither carefully designed tasks nor established
evaluation methods for mathematical content access. With this background
in mind, the NTCIR Math Task pushes forward the establishment of new
challenges in mathematical information retrieval by providing shared
datasets and a common evaluation platform to interested researchers.


### TASKS ###

Math Retrieval Subtask:

Given a provided dataset, systems must return a ranked list of search
results. There are three different search scenarios:

    Formula Search (automated)
        Search for formula queries within the formulae database of the
used dataset
    Full-Text Search (automated)
        Search the document collection using formula queries. Queries
are combinations of keywords and formulae.
    Open Information Retrieval (semi-automated)
        Search the document collection using free textual queries.

Math Understanding Subtask:

Given a scientific paper with its MathML representation, extract all the
natural language definitions for the mathematical objects.


### IMPORTANT DATES ###

May, 2012	Initial dataset and example topics released
Early-mid October, 2012	Topic release for Open IR search type
Late October, 2012	Topic release for Formula, Full-text search types,
and Math understanding subtask
Early November, 2012	Submissions due for Formula, Full-text search
types, and Math Understanding subtask
Mid November, 2012	Submissions due for Open IR search type
February, 2013	Evaluation results released
March, 2013	Draft papers for NTCIR-10 Proceedings due
May, 2013	Camera ready for NTCIR-10 Proceedings due
June 18-21 2013	NTCIR-10 Meeting

### DATASETS ###

Participants need to download the datasets for Math Retrieval Subtask
and/or Math Understanding Subtask on the NTCIR Pilot Task: Math Task
website.
Dataset and example topics are available upon request. Please contact
NTCIR-math organizers :
ntcir10adm-math (at) nii.ac.jp


### HOW TO PARTICIPATE ###

Task registration page is available on the NTCIR-10 website.
http://research.nii.ac.jp/ntcir/ntcir-10/howto.html

Please note that task participants must write a paper for the NTCIR-10
meeting (unrefereed) and have at least one representative attending the
meeting at NII, Tokyo in June 2013.


### ORGANISERS ###

Akiko Aizawa, National Institute of Informatics, Japan
Michael Kohlhase, Jacobs University Bremen
Iadh Ounis, University of Glasgow

###  QUESTIONS? PLEASE CONTACT ###

ntcir10adm-math (at) nii.ac.jp

### LINKS ###

NTCIR Pilot Task: Math Task website
http://ntcir-math.nii.ac.jp/