Peopleviews: Human Computation for Constraint-Based Recommendation

W I S S E N T E C H N I K L E I D E N S C H A F T

contact: thomas.ulz@student.tugraz.at // http://www.ist.tugraz.at/

PeopleViews: Human Computation forConstraint-Based Recommendation

Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz,Stefan Reiterer, Martin Stettinger, Institute for Software Technology,Graz University of Technology

2 Contents

1. Motivation

2. PeopleViews

3. Recommendation approach

4. Evaluation

5. Ongoing and future work

Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger

Motivation

Constraint-based Recommendation

Specific type of knowledge based recommendation

Relies on a predefined set of constraints

Rankings determined by utility function

Why constraint-based recommendation?

Suitable for complex item domainsPossible to ”explain“ recommendationsDiagnoses for too strict requirements

Motivation

Knowledge acquisition bottleneck

Only a few KnowledgeEngineers

Possibly a lot of userswith item knowledge

Idea: enable users tocontribute to knowledgebases

Are users willing tocontribute to knowledgebases?

End Users

KnowledgeEngineers

Item Knowledge Knowledge Base

Motivation

How willing are end users to contribute?

N=161, 111 would be willing to contribute

[Felfernig et al., CrowdRec 2014]

PeopleViews

Short-term tasks (Micro-tasks)

Domain experts perform short-term knowledgeengineering tasks they are much better incompared to knowledge engineers.

Potential advantages

Less effort related to recommendation knowledgebase development and maintenanceFewer erroneous constraintsSignificantly higher degree of scalability

PeopleViews

PeopleViews - Knowledge base

Product attributes

”Facts“ about items, e.g. sensor size of a cameraDefined when item is added to knowledge base

User attributes

Perceived differently by users, e.g. a camerasfield of applicationDefined by users in micro tasks

Support

Support of item for specific {user, product}attribute value

PeopleViews

PeopleViews - Features

Users are able to:

Define new knowledge bases

Create new recommendation domainAdd items to existing domainsEvaluate existing items

”Answer“ micro tasks

Use existing knowledge bases to getrecommendations

PeopleViews

Definition of knowledge bases

Product attributes

attribute question to user domain similaritymetric

sensorsize Preferred sensorsize?

{fullframe, APS-C,MFT, 1“, 2/3“} EIB

max-shutterspeed

Required max.shutter speed?

{1/4000, 1/6000,1/8000, 1/16000} LIB

maxISO Required max.ISO sensitivity?

{6400, 12800,25600} MIB

price Max. price? integer LIB

PeopleViews

User attributes

attribute choice type question to user domainusertype multiple Suited for whom? { beginner, amateur, expert }

application single Preferred { sport, architecture, macro,application? landscape, portrait }

usability single Minimum accepted { average, high, very high }usability?

PeopleViews

Micro-tasks

choice type: multipleAlexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger

PeopleViews

Micro-tasks

choice type: single

PeopleViews

Micro-tasks

choose item, single attributeAlexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger

PeopleViews

Recommendation view

PeopleViews

Recommendation view - Item details

Recommendation approach

Recommendation approach in PeopleViews

User attributes

support(Φ,u, v) =

∑s(Φ,u, v)

|s(Φ,u, v)|· |s(Φ,u, v)||s(Φ,u)|

symbol meaningΦ itemu user attribute u ∈ Up product attributev {user, product} attribute value

s(Φ, u, v) support specified by user

Product attributes

support(Φ,p, v) =

1if v = val(Φ, p), 0 otherwise EIB

1− |v - val(Φ, p)|max(Φ, p) - min(Φ, p)

val(Φ, p)-min(Φ, p)max(Φ, p) - min(Φ, p)

max(Φ, p)-val(Φ, p)max(Φ, p) - min(Φ ,p)

Selection of recommendation-relevant items

f(Φ) =∧u∈U

u ∈ values(Φ, u) ∪ {noval} → include(Φ)

Ranking items by their utility

utility(Φ, REQ) =∑

a=v∈REQ

support(Φ, a, v) ·w(a)

Evaluation

Evaluation of recommender algorithms

Collect data using WeeVis (http://www.weevis.org)

Canon DSLR recommender

16 items, 7 attributes (27 possible ”answers“)

Users defined their requirements and selected bestmatching camera

356 unique sessions

1 out of N ”training and evaluation“

Is desired item in top n recommended items?

Evaluation

Comparison to other approaches

1 1.5 2 2.5 3 3.5 4 4.5 50

ion in %

Learned Weights

Support Values

Popularity Based

Random

Ongoing and future work

Recommendation approaches

Implementation and evaluation of furtherapproaches; diagnoses and repair

Micro-task scheduling

Automatically assign micro-tasks to users, usinga content-based approach

Quality assurance

Improve dataset quality and prevent manipulation

Thank you!

Peopleviews: Human Computation for Constraint-Based Recommendation

Data & Analytics

Transcript of Peopleviews: Human Computation for Constraint-Based Recommendation

Non-binary constraints Toby Walsh Cork Constraint Computation Center tw@4c.ucc.ie.

Search for satisfaction Toby Walsh Cork Constraint Computation Center tw@4c.ucc.ie.

DATA GATHERING AND ROUTE OPTIMIZATION PROTOCOL …...Limited computation power and memory size is another constraint that ... rooting or aggregation protocols, they ... often preset

Group Recommendation: Semantics and Efficiency09.pdf · Efficient Recommendation Computation i1,2 i2,2 i3,2 i4,2 i4,2 i3,2 i1,0 i2,0 IL 1 IL 2 - Ordina in maniera decrescente IL -

Boundary-based Correspondence Computation … Correspondence Computation Using the Topology Constraint T. Rachidi* and L. Spacek Department of Computer Science, University of Essex,

Context and Intention in Ontologies Rick Wallace & Tabbasum Naz Cork Constraint Computation Centre University College Cork Cork, Ireland ARCOE 2010, Lisbon.

A Timed Concurrent Constraint Language · PDF fileInformation and Computation 161, 45 83 (2000) A Timed Concurrent Constraint Language F. S. de Boer Department of Computer Science,

Income Computation & Disclosure Standards · Recommendation for proper implementation of the Income Computation & Disclosure Standards (‘IDS’) December 2015 American Chamber of

A Simple Inference Framework for Connecting the Dots Jacob Feldman, PhD OpenRules, Inc. Cork Constraint Computation Centre .

A Logic of Partially Satisfied Constraints Nic Wilson Cork Constraint Computation Centre Computer Science, UCC.

Capacity-aware Sequential Recommendations · decouple constraint satisfaction from sequential recommendation policies, resulting in algorithms which issue recommendations to thousands

Constraint Conjunction versus Grounded Constraint ...

12 October 2006 Foundations of Logic and Constraint Programming 1 Operational Semantics An overview The language of programs The computation mechanism.

Fundamental Matrix Computation: Theory and …kanatani/papers/thefund.pdfFundamental Matrix Computation: Theory and Practice ... solution that satisﬁes the rank constraint automatically

Computation of Limit Cycles for Uncertain Nonlinear Fractional-order Systems using Interval Constraint Propagation P. S. V. Nataraj & Rambabu Kalla Systems.

Non-binary constraints: modelling Toby Walsh Cork Constraint Computation Center.

Constraint Programming (CP) - SINTEF · Constraint Programming History Constraint ... Constraint Programming (CP) ... Constraint Programming one of the basic technologies for constructing

Constraint Propagation: The Heart of Constraint Programming

Constraint Solving and Constraint Satisfaction Search

Constraint Programming for Supply Chain Management J. Christopher Beck Cork Constraint Computation Centre (4C) c.beck@4c.ucc.ie SCM Information Day, Nov.