Inter-sentence features and thresholded minimum error rate training

NAIST at CLEF 2013 QA4MRE

Philip Arthur, Graham Neubig, Sakriani Sakti, Tomoki Toda, Satoshi Nakamura

Research output: Contribution to journalConference articleResearchpeer-review

1 Citation (Scopus)

Abstract

This paper describes the Nara Institute of Science and Technology's system for the main task of CLEF 2013 QA4MRE. The core of the system is a log linear scoring model that couples both intra and intersentence features. Each of the features receives an input of a candidate answer, question, and document, and uses these to assign a score according to some criterion. We use minimum error rate training (MERT) to train the weights of the model and also propose a novel method for MERT with the addition of a threshold that defines the certainty with which we must answer questions. The system received a score of 28% c@1 on main questions and 33% c@1 when considering auxiliary questions on the CLEF 2013 evaluation.

Original languageEnglish
JournalCEUR Workshop Proceedings
Volume1179
Publication statusPublished - 1 Jan 2013
Event2013 Cross Language Evaluation Forum Conference, CLEF 2013 - Valencia, Spain
Duration: 23 Sep 201326 Sep 2013

Keywords

  • Discriminative learning
  • Inter-sentence features
  • Linear feature model
  • Machine reading
  • Minimum error rate training
  • Question answering

Cite this

Arthur, Philip ; Neubig, Graham ; Sakti, Sakriani ; Toda, Tomoki ; Nakamura, Satoshi. / Inter-sentence features and thresholded minimum error rate training : NAIST at CLEF 2013 QA4MRE. In: CEUR Workshop Proceedings. 2013 ; Vol. 1179.
@article{fbfd623fd3cd4d77aebe19a57238f478,
title = "Inter-sentence features and thresholded minimum error rate training: NAIST at CLEF 2013 QA4MRE",
abstract = "This paper describes the Nara Institute of Science and Technology's system for the main task of CLEF 2013 QA4MRE. The core of the system is a log linear scoring model that couples both intra and intersentence features. Each of the features receives an input of a candidate answer, question, and document, and uses these to assign a score according to some criterion. We use minimum error rate training (MERT) to train the weights of the model and also propose a novel method for MERT with the addition of a threshold that defines the certainty with which we must answer questions. The system received a score of 28{\%} c@1 on main questions and 33{\%} c@1 when considering auxiliary questions on the CLEF 2013 evaluation.",
keywords = "Discriminative learning, Inter-sentence features, Linear feature model, Machine reading, Minimum error rate training, Question answering",
author = "Philip Arthur and Graham Neubig and Sakriani Sakti and Tomoki Toda and Satoshi Nakamura",
year = "2013",
month = "1",
day = "1",
language = "English",
volume = "1179",
journal = "CEUR Workshop Proceedings",
issn = "1613-0073",
publisher = "Rheinisch-Westfaelische Technische Hochschule Aachen * Lehrstuhl Informatik V",

}

Inter-sentence features and thresholded minimum error rate training : NAIST at CLEF 2013 QA4MRE. / Arthur, Philip; Neubig, Graham; Sakti, Sakriani; Toda, Tomoki; Nakamura, Satoshi.

In: CEUR Workshop Proceedings, Vol. 1179, 01.01.2013.

Research output: Contribution to journalConference articleResearchpeer-review

TY - JOUR

T1 - Inter-sentence features and thresholded minimum error rate training

T2 - NAIST at CLEF 2013 QA4MRE

AU - Arthur, Philip

AU - Neubig, Graham

AU - Sakti, Sakriani

AU - Toda, Tomoki

AU - Nakamura, Satoshi

PY - 2013/1/1

Y1 - 2013/1/1

N2 - This paper describes the Nara Institute of Science and Technology's system for the main task of CLEF 2013 QA4MRE. The core of the system is a log linear scoring model that couples both intra and intersentence features. Each of the features receives an input of a candidate answer, question, and document, and uses these to assign a score according to some criterion. We use minimum error rate training (MERT) to train the weights of the model and also propose a novel method for MERT with the addition of a threshold that defines the certainty with which we must answer questions. The system received a score of 28% c@1 on main questions and 33% c@1 when considering auxiliary questions on the CLEF 2013 evaluation.

AB - This paper describes the Nara Institute of Science and Technology's system for the main task of CLEF 2013 QA4MRE. The core of the system is a log linear scoring model that couples both intra and intersentence features. Each of the features receives an input of a candidate answer, question, and document, and uses these to assign a score according to some criterion. We use minimum error rate training (MERT) to train the weights of the model and also propose a novel method for MERT with the addition of a threshold that defines the certainty with which we must answer questions. The system received a score of 28% c@1 on main questions and 33% c@1 when considering auxiliary questions on the CLEF 2013 evaluation.

KW - Discriminative learning

KW - Inter-sentence features

KW - Linear feature model

KW - Machine reading

KW - Minimum error rate training

KW - Question answering

UR - http://www.scopus.com/inward/record.url?scp=84922021519&partnerID=8YFLogxK

M3 - Conference article

VL - 1179

JO - CEUR Workshop Proceedings

JF - CEUR Workshop Proceedings

SN - 1613-0073

ER -