A Bayesian Concept Learning Approach to Crowdsourcing

Paolo Renato Viappiani; Sandra Zilles; Howard J. Hamilton; Craiug Boutilier

A Bayesian Concept Learning Approach to Crowdsourcing

Paolo Renato Viappiani, Sandra Zilles, Howard J. Hamilton, Craiug Boutilier

Department of Computer Science

Research output: Contribution to journal › Conference article in Journal › Research › peer-review

Abstract

We develop a Bayesian approach to concept learning for crowdsourcing applications. A probabilistic belief over possible concept definitions is maintained and updated according to (noisy) observations from experts, whose behaviors are modeled using discrete types. We propose recommendation techniques, inference methods, and query selection
strategies to assist a user charged with choosing a configuration that satisfies some (partially known) concept. Our model is able to simultaneously learn
the concept definition and the types of the experts. We evaluate our model with simulations, showing that our Bayesian strategies are effective even in large concept spaces with many uninformative experts.

Original language	English
Journal	CEUR Workshop Proceedings
Volume	756
Number of pages	8
ISSN	1613-0073
Publication status	Published - 2011
Event	INTELLIGENT TECHNIQUES FOR WEB PERSONALIZATION & RECOMMENDER SYSTEMS (ITWP'11) - Barcelona, Spain Duration: 16 Jul 2011 → … Conference number: 9

Workshop

Workshop	INTELLIGENT TECHNIQUES FOR WEB PERSONALIZATION & RECOMMENDER SYSTEMS (ITWP'11)
Number	9
Country/Territory	Spain
City	Barcelona
Period	16/07/2011 → …

Access to Document

http://ceur-ws.org/

AUB Link

Search for the material in Aalborg University Library's search engine

Cite this

@inproceedings{b2d711eeb9ac4bbda3df818193e672f3,

title = "A Bayesian Concept Learning Approach to Crowdsourcing",

abstract = "We develop a Bayesian approach to concept learning for crowdsourcing applications. A probabilistic belief over possible concept definitions is maintained and updated according to (noisy) observations from experts, whose behaviors are modeled using discrete types. We propose recommendation techniques, inference methods, and query selectionstrategies to assist a user charged with choosing a configuration that satisfies some (partially known) concept. Our model is able to simultaneously learnthe concept definition and the types of the experts. We evaluate our model with simulations, showing that our Bayesian strategies are effective even in large concept spaces with many uninformative experts.",

author = "Viappiani, {Paolo Renato} and Sandra Zilles and Hamilton, {Howard J.} and Craiug Boutilier",

year = "2011",

language = "English",

volume = "756",

journal = "CEUR Workshop Proceedings",

issn = "1613-0073",

publisher = "CEUR Workshop Proceedings",

note = "INTELLIGENT TECHNIQUES FOR WEB PERSONALIZATION & RECOMMENDER SYSTEMS (ITWP'11) ; Conference date: 16-07-2011",

}

TY - GEN

T1 - A Bayesian Concept Learning Approach to Crowdsourcing

AU - Viappiani, Paolo Renato

AU - Zilles, Sandra

AU - Hamilton, Howard J.

AU - Boutilier, Craiug

N1 - Conference code: 9

PY - 2011

Y1 - 2011

N2 - We develop a Bayesian approach to concept learning for crowdsourcing applications. A probabilistic belief over possible concept definitions is maintained and updated according to (noisy) observations from experts, whose behaviors are modeled using discrete types. We propose recommendation techniques, inference methods, and query selectionstrategies to assist a user charged with choosing a configuration that satisfies some (partially known) concept. Our model is able to simultaneously learnthe concept definition and the types of the experts. We evaluate our model with simulations, showing that our Bayesian strategies are effective even in large concept spaces with many uninformative experts.

AB - We develop a Bayesian approach to concept learning for crowdsourcing applications. A probabilistic belief over possible concept definitions is maintained and updated according to (noisy) observations from experts, whose behaviors are modeled using discrete types. We propose recommendation techniques, inference methods, and query selectionstrategies to assist a user charged with choosing a configuration that satisfies some (partially known) concept. Our model is able to simultaneously learnthe concept definition and the types of the experts. We evaluate our model with simulations, showing that our Bayesian strategies are effective even in large concept spaces with many uninformative experts.

M3 - Conference article in Journal

SN - 1613-0073

VL - 756

JO - CEUR Workshop Proceedings

JF - CEUR Workshop Proceedings

T2 - INTELLIGENT TECHNIQUES FOR WEB PERSONALIZATION & RECOMMENDER SYSTEMS (ITWP'11)

Y2 - 16 July 2011

ER -

A Bayesian Concept Learning Approach to Crowdsourcing

Abstract

Workshop

Access to Document

AUB Link

Fingerprint

Cite this