Cross-domain answer ranking using importance sampling

Anders Trærup Johannsen; Anders Søgaard

Cross-domain answer ranking using importance sampling

Anders Trærup Johannsen, Anders Søgaard

LUKKET: Center for Sprogteknologi

Abstract

We consider the problem of learning how to rank answers across domains in community question answering using stylistic features. Our main contribution is an importance sampling technique for selecting training data per answer thread. Our approach is evaluated across 30 community sites and shown to be significantly better than random sampling. We show that the most useful features in our model relate to answer length and overlap with question.

Originalsprog	Engelsk
Titel	The 6th International Joint Conference on Natural Language Processing (IJCNLP)
Forlag	Association for Computational Linguistics
Publikationsdato	2013
Sider	987-991
ISBN (Elektronisk)	978-4-9907348-0-0
Status	Udgivet - 2013
Begivenhed	The 6th International Joint Conference on Natural Language Processing (IJCNLP): IJCNLP - Nagoya, Japan Varighed: 14 okt. 2013 → 18 maj 2014 Konferencens nummer: 6

Konference

Konference	The 6th International Joint Conference on Natural Language Processing (IJCNLP)
Nummer	6
Land/Område	Japan
By	Nagoya
Periode	14/10/2013 → 18/05/2014

Citationsformater

@inproceedings{858ba02c80a447dd9f432a502a6e57a7,

title = "Cross-domain answer ranking using importance sampling",

abstract = "We consider the problem of learning how to rank answers across domains in community question answering using stylistic features. Our main contribution is an importance sampling technique for selecting training data per answer thread. Our approach is evaluated across 30 community sites and shown to be significantly better than random sampling. We show that the most useful features in our model relate to answer length and overlap with question.",

author = "Johannsen, {Anders Tr{\ae}rup} and Anders S{\o}gaard",

year = "2013",

language = "English",

pages = "987--991",

booktitle = "The 6th International Joint Conference on Natural Language Processing (IJCNLP)",

publisher = "Association for Computational Linguistics",

note = "The 6th International Joint Conference on Natural Language Processing (IJCNLP) ; Conference date: 14-10-2013 Through 18-05-2014",

}

TY - GEN

T1 - Cross-domain answer ranking using importance sampling

AU - Johannsen, Anders Trærup

AU - Søgaard, Anders

N1 - Conference code: 6

PY - 2013

Y1 - 2013

N2 - We consider the problem of learning how to rank answers across domains in community question answering using stylistic features. Our main contribution is an importance sampling technique for selecting training data per answer thread. Our approach is evaluated across 30 community sites and shown to be significantly better than random sampling. We show that the most useful features in our model relate to answer length and overlap with question.

AB - We consider the problem of learning how to rank answers across domains in community question answering using stylistic features. Our main contribution is an importance sampling technique for selecting training data per answer thread. Our approach is evaluated across 30 community sites and shown to be significantly better than random sampling. We show that the most useful features in our model relate to answer length and overlap with question.

M3 - Article in proceedings

SP - 987

EP - 991

BT - The 6th International Joint Conference on Natural Language Processing (IJCNLP)

PB - Association for Computational Linguistics

T2 - The 6th International Joint Conference on Natural Language Processing (IJCNLP)

Y2 - 14 October 2013 through 18 May 2014

ER -

Cross-domain answer ranking using importance sampling

Abstract

Konference

Fingeraftryk

Citationsformater