An Ensemble Model with Ranking for Social Dialogue

Ioannis Papaioannou; Amanda Cercas Curry; Jose Part; Igor Shalyminov; Xu Xinnuo; Yanchao Yu; Ondrej Dusek; Verena Rieser; Oliver Lemon

An Ensemble Model with Ranking for Social Dialogue

Ioannis Papaioannou, Amanda Cercas Curry, Jose Part, Igor Shalyminov, Xu Xinnuo, Yanchao Yu, Ondrej Dusek, Verena Rieser, Oliver Lemon

Research output: Contribution to conference › Paper › peer-review

Abstract

Open-domain social dialogue is one of the long-standing goals of Artificial Intelligence. This year, the Amazon Alexa Prize challenge was announced for the first time, where real customers get to rate systems developed by leading universities worldwide. The aim of the challenge is to converse “coherently and engagingly with humans on popular topics for 20 minutes”. We describe our Alexa Prize system (called ‘Alana’) consisting of an ensemble of bots, combining rule-based and machine learning systems, and using a contextual ranking mechanism to choose a system response. The ranker was trained on real user feedback received during the competition, where we address the problem of how to train on the noisy and sparse feedback obtained during the competition.

Original language	English
Publication status	Published - 2017
Event	NIPS 2017 Conversational AI Workshop - Long Beach Convention Center, Long Beach, United States Duration: 8 Dec 2017 → 8 Dec 2017

Workshop

Workshop	NIPS 2017 Conversational AI Workshop
Country/Territory	United States
City	Long Beach
Period	8/12/17 → 8/12/17

Access to Document

https://arxiv.org/abs/1712.07558

Cite this

@conference{87af09cddecf40a2a7a3af1f5daf54ee,

title = "An Ensemble Model with Ranking for Social Dialogue",

abstract = "Open-domain social dialogue is one of the long-standing goals of Artificial Intelligence. This year, the Amazon Alexa Prize challenge was announced for the first time, where real customers get to rate systems developed by leading universities worldwide. The aim of the challenge is to converse “coherently and engagingly with humans on popular topics for 20 minutes”. We describe our Alexa Prize system (called {\textquoteleft}Alana{\textquoteright}) consisting of an ensemble of bots, combining rule-based and machine learning systems, and using a contextual ranking mechanism to choose a system response. The ranker was trained on real user feedback received during the competition, where we address the problem of how to train on the noisy and sparse feedback obtained during the competition.",

author = "Ioannis Papaioannou and {Cercas Curry}, Amanda and Jose Part and Igor Shalyminov and Xu Xinnuo and Yanchao Yu and Ondrej Dusek and Verena Rieser and Oliver Lemon",

year = "2017",

language = "English",

note = "NIPS 2017 Conversational AI Workshop ; Conference date: 08-12-2017 Through 08-12-2017",

}

TY - CONF

T1 - An Ensemble Model with Ranking for Social Dialogue

AU - Papaioannou, Ioannis

AU - Cercas Curry, Amanda

AU - Part, Jose

AU - Shalyminov, Igor

AU - Xinnuo, Xu

AU - Yu, Yanchao

AU - Dusek, Ondrej

AU - Rieser, Verena

AU - Lemon, Oliver

PY - 2017

Y1 - 2017

N2 - Open-domain social dialogue is one of the long-standing goals of Artificial Intelligence. This year, the Amazon Alexa Prize challenge was announced for the first time, where real customers get to rate systems developed by leading universities worldwide. The aim of the challenge is to converse “coherently and engagingly with humans on popular topics for 20 minutes”. We describe our Alexa Prize system (called ‘Alana’) consisting of an ensemble of bots, combining rule-based and machine learning systems, and using a contextual ranking mechanism to choose a system response. The ranker was trained on real user feedback received during the competition, where we address the problem of how to train on the noisy and sparse feedback obtained during the competition.

AB - Open-domain social dialogue is one of the long-standing goals of Artificial Intelligence. This year, the Amazon Alexa Prize challenge was announced for the first time, where real customers get to rate systems developed by leading universities worldwide. The aim of the challenge is to converse “coherently and engagingly with humans on popular topics for 20 minutes”. We describe our Alexa Prize system (called ‘Alana’) consisting of an ensemble of bots, combining rule-based and machine learning systems, and using a contextual ranking mechanism to choose a system response. The ranker was trained on real user feedback received during the competition, where we address the problem of how to train on the noisy and sparse feedback obtained during the competition.

M3 - Paper

T2 - NIPS 2017 Conversational AI Workshop

Y2 - 8 December 2017 through 8 December 2017

ER -

An Ensemble Model with Ranking for Social Dialogue

Abstract

Workshop

Access to Document

Fingerprint

Cite this