SlideShare una empresa de Scribd logo
1 de 14
Descargar para leer sin conexión
© Fraunhofer IDMT
T. Köllmer†, E. Berndl‡, T. Weißgerber‡, P. Aichroth† , H. Kosch‡
A Workflow for Cross Media Recommendations
based on Linked Data Analysis
† Fraunhofer IDMT ‡ University of Passau
Thomas Köllmer
thomas.koellmer@idmt.fraunhofer.de
© Fraunhofer IDMT 2
The MICO Project
 http://www.mico-project.eu
 FP7 STREP, GA# 610480
 7 Partners: Salzburg Research, Fraunhofer IDMT, Oxford University,
University of Passau, Umeå University, Zaizi, InsideOut10
 Duration: 36 Months (end: 10/2016)
 Goal of MICO is to develop a platform to analyse “media in context” …
 ...by orchestrating different content extraction tools that can work
simultaneously or in a sequence and …..
 ...provide valuable metadata to third party applications.
© Fraunhofer IDMT 3
Cross Media Recommendations in MICO
 The Recommendation Problem
 “Providing suggestions for items to be of use to a user”
 Broad Application domains:
 e-Commerce, Travel, Multimedia, News, …
 Two main data sources:
 User behaviour data (Collaborative Filtering)
 Item Metadata (Content based recommendation)
© Fraunhofer IDMT 4
Cross Media Recommendations in MICO
 Media Recommendation
 “Classic” Recommendation problems (Music-Recommendation, Netflix
Challenge)
 Mostly based on collaborative filtering, but also content analysis
 Cross Media Recommendation
 Cross Media ≠ Cross Modal
 Multiple Input
 Media Items + Associated Media context, e.g., commentary
 Multiple Output
 Not necessarily more of the same thing
© Fraunhofer IDMT 5
Editor Support Usecase
As an editor, while I create or edit articles using WordPress, I want to
automatically
get related articles and videos that I might link to the article.
 Project Use Case: Recommendations for Greenpeace Magazine Italia
 Articles
 Videos
 User behaviour data
© Fraunhofer IDMT 6
Editor Support Usecase
As an editor, while I create or edit articles using WordPress, I want to
automatically
get related articles and videos that I might link to the article.
 How can Mico Help?
 Articles: NER components for content spotting
 Videos
 Text to speech + NER
 Metadata Analysis
 User behaviour data
 Generic collaborative filtering component (prediction.io)
© Fraunhofer IDMT 7
Chat Analysis Usecase
As a zooniverse user, I want get more information that helps me during the
current classification task, e.g., by the recommendation of similar subjects
or training data.
 Input:
 Results from animal detection
 Sentiment and competence classification of talks
 NER for animal spotting
 Output:
 Classification Hints
 Hints for moderators / experts
© Fraunhofer IDMT 8
Processing Workflow for Editor Support Use Case
1. Crawler feeds media items to platform
e.g., videos from Greenpeace channel, news articles, internal archive….
2. Analysis results are stored as Linked Data
Currently: using Apache Marmotta as triplestore
© Fraunhofer IDMT 9
Processing Workflow for Editor Support Use Case
3. Item gets into Focus (someone is writing an article on a specific topic)
4. Focused item is fed to the same platform, using customized pipelines
5. Results are processed by LD matching, i.e Recommendation component..
6. … enriched by stored annotations …
7. And returned to editing platform
© Fraunhofer IDMT 10
Recommendation Workflow in MICO
© Fraunhofer IDMT 11
Ontology Matching
 Semantic distance of entities important for analysis
 Animal detection: “I see a bird on the tree”, “Yes it’s definitely a
sparrow”
 Topic detection: Wind energy, sustainability, renewables, oil, …
 Non-Mico:
 Genres! (e.g., currently 1400 genres on Spotify…)
 Not less controversial: movie genres
© Fraunhofer IDMT 12
Evaluation
 How to evaluate a recommender system?
 Forget about precision and recall!
 Observing user behaviour
 Online experiment
 Offline: extrapolating past user behavior
 Lab Studies / Questionnaire
 Open problem: standardized + comparable testing of recommender
systems
© Fraunhofer IDMT 13
Conclusions
 Proposed setting tries to formalize a certain subset of recommendation
problems: Cross media recommendation
 Focused on Media Items
 Profits from Content Analysis
 New items, have a high contextual dependency on existing items, that
will be exploited for recommendation
 Work in Progress:
 Stable release of Mico-platform and extractors: right now
 PoC recommendation code will be integrated in the platform
 Demo Services: available in June
 Incorporate feedback into recommendation
© Fraunhofer IDMT
T. Köllmer†, E. Berndl‡, T. Weißgerber‡, P. Aichroth† , H. Kosch‡
A Workflow for Cross Media Recommendations
based on Linked Data Analysis
† Fraunhofer IDMT ‡ University of Passau
Thank you!
Questions?

Más contenido relacionado

Similar a Lime recommendation

Research Careers in Applied Computer Science
Research Careers in Applied Computer ScienceResearch Careers in Applied Computer Science
Research Careers in Applied Computer ScienceChristoph Lange
 
Summer school bz_fp7research_20100708
Summer school bz_fp7research_20100708Summer school bz_fp7research_20100708
Summer school bz_fp7research_20100708Sandro D'Elia
 
Go open2010 sde_20100417
Go open2010 sde_20100417Go open2010 sde_20100417
Go open2010 sde_20100417Sandro D'Elia
 
Fraunhofer iais audio mining - automatic metadata gereration of audio streams...
Fraunhofer iais audio mining - automatic metadata gereration of audio streams...Fraunhofer iais audio mining - automatic metadata gereration of audio streams...
Fraunhofer iais audio mining - automatic metadata gereration of audio streams...FIAT/IFTA
 
GoOpen 2010: Sandro D'Elia
GoOpen 2010: Sandro D'EliaGoOpen 2010: Sandro D'Elia
GoOpen 2010: Sandro D'EliaFriprogsenteret
 
Better Software, Better Research
Better Software, Better ResearchBetter Software, Better Research
Better Software, Better ResearchCarole Goble
 
LEARNTEC 2018 - HANDELkompetent - Enabling Adaptive Learning Pathways by Situ...
LEARNTEC 2018 - HANDELkompetent - Enabling Adaptive Learning Pathways by Situ...LEARNTEC 2018 - HANDELkompetent - Enabling Adaptive Learning Pathways by Situ...
LEARNTEC 2018 - HANDELkompetent - Enabling Adaptive Learning Pathways by Situ...Sebastian Opriel
 
NLP on Hadoop: A Distributed Framework for NLP-Based Keyword and Keyphrase Ex...
NLP on Hadoop: A Distributed Framework for NLP-Based Keyword and Keyphrase Ex...NLP on Hadoop: A Distributed Framework for NLP-Based Keyword and Keyphrase Ex...
NLP on Hadoop: A Distributed Framework for NLP-Based Keyword and Keyphrase Ex...Paolo Nesi
 
Open and Collaborative Software for Digital Pathology
Open and Collaborative Software for Digital Pathology Open and Collaborative Software for Digital Pathology
Open and Collaborative Software for Digital Pathology William Baird
 
MICO — Towards Contextual Media Analysis
MICO — Towards Contextual Media AnalysisMICO — Towards Contextual Media Analysis
MICO — Towards Contextual Media AnalysisThomas Kurz
 
Hamburg Requirements Engineering Symposium
Hamburg Requirements Engineering SymposiumHamburg Requirements Engineering Symposium
Hamburg Requirements Engineering SymposiumSupersede
 
Wageningen phenotype meeting
Wageningen phenotype meetingWageningen phenotype meeting
Wageningen phenotype meetingthehyve
 
Overview of methodologies
Overview of methodologiesOverview of methodologies
Overview of methodologiesMickael Pero
 
Frankfurt Big Data Lab & Refugee Projeect
Frankfurt Big Data Lab & Refugee ProjeectFrankfurt Big Data Lab & Refugee Projeect
Frankfurt Big Data Lab & Refugee ProjeectGoethe Univeristy
 
Jonathan bright - collecting social media data with the python programming la...
Jonathan bright - collecting social media data with the python programming la...Jonathan bright - collecting social media data with the python programming la...
Jonathan bright - collecting social media data with the python programming la...oiisdp
 
ISCRAM Summer School lecture Prof. Ralph Morelli
ISCRAM Summer School lecture Prof. Ralph MorelliISCRAM Summer School lecture Prof. Ralph Morelli
ISCRAM Summer School lecture Prof. Ralph Morelliwjm
 
OpenNTF Overview DanNotes 11/23/11
OpenNTF Overview DanNotes 11/23/11OpenNTF Overview DanNotes 11/23/11
OpenNTF Overview DanNotes 11/23/11Niklas Heidloff
 
Data management plans – EUDAT Best practices and case study | www.eudat.eu
Data management plans – EUDAT Best practices and case study | www.eudat.euData management plans – EUDAT Best practices and case study | www.eudat.eu
Data management plans – EUDAT Best practices and case study | www.eudat.euEUDAT
 

Similar a Lime recommendation (20)

Research Careers in Applied Computer Science
Research Careers in Applied Computer ScienceResearch Careers in Applied Computer Science
Research Careers in Applied Computer Science
 
Summer school bz_fp7research_20100708
Summer school bz_fp7research_20100708Summer school bz_fp7research_20100708
Summer school bz_fp7research_20100708
 
Go open2010 sde_20100417
Go open2010 sde_20100417Go open2010 sde_20100417
Go open2010 sde_20100417
 
Fraunhofer iais audio mining - automatic metadata gereration of audio streams...
Fraunhofer iais audio mining - automatic metadata gereration of audio streams...Fraunhofer iais audio mining - automatic metadata gereration of audio streams...
Fraunhofer iais audio mining - automatic metadata gereration of audio streams...
 
GoOpen 2010: Sandro D'Elia
GoOpen 2010: Sandro D'EliaGoOpen 2010: Sandro D'Elia
GoOpen 2010: Sandro D'Elia
 
Better Software, Better Research
Better Software, Better ResearchBetter Software, Better Research
Better Software, Better Research
 
LEARNTEC 2018 - HANDELkompetent - Enabling Adaptive Learning Pathways by Situ...
LEARNTEC 2018 - HANDELkompetent - Enabling Adaptive Learning Pathways by Situ...LEARNTEC 2018 - HANDELkompetent - Enabling Adaptive Learning Pathways by Situ...
LEARNTEC 2018 - HANDELkompetent - Enabling Adaptive Learning Pathways by Situ...
 
NLP on Hadoop: A Distributed Framework for NLP-Based Keyword and Keyphrase Ex...
NLP on Hadoop: A Distributed Framework for NLP-Based Keyword and Keyphrase Ex...NLP on Hadoop: A Distributed Framework for NLP-Based Keyword and Keyphrase Ex...
NLP on Hadoop: A Distributed Framework for NLP-Based Keyword and Keyphrase Ex...
 
2016 nov-ieee-sdn-wiki
2016 nov-ieee-sdn-wiki2016 nov-ieee-sdn-wiki
2016 nov-ieee-sdn-wiki
 
Open and Collaborative Software for Digital Pathology
Open and Collaborative Software for Digital Pathology Open and Collaborative Software for Digital Pathology
Open and Collaborative Software for Digital Pathology
 
MICO — Towards Contextual Media Analysis
MICO — Towards Contextual Media AnalysisMICO — Towards Contextual Media Analysis
MICO — Towards Contextual Media Analysis
 
Hamburg Requirements Engineering Symposium
Hamburg Requirements Engineering SymposiumHamburg Requirements Engineering Symposium
Hamburg Requirements Engineering Symposium
 
Wageningen phenotype meeting
Wageningen phenotype meetingWageningen phenotype meeting
Wageningen phenotype meeting
 
Overview of methodologies
Overview of methodologiesOverview of methodologies
Overview of methodologies
 
Frankfurt Big Data Lab & Refugee Projeect
Frankfurt Big Data Lab & Refugee ProjeectFrankfurt Big Data Lab & Refugee Projeect
Frankfurt Big Data Lab & Refugee Projeect
 
Jonathan bright - collecting social media data with the python programming la...
Jonathan bright - collecting social media data with the python programming la...Jonathan bright - collecting social media data with the python programming la...
Jonathan bright - collecting social media data with the python programming la...
 
ISCRAM Summer School lecture Prof. Ralph Morelli
ISCRAM Summer School lecture Prof. Ralph MorelliISCRAM Summer School lecture Prof. Ralph Morelli
ISCRAM Summer School lecture Prof. Ralph Morelli
 
OpenNTF Overview DanNotes 11/23/11
OpenNTF Overview DanNotes 11/23/11OpenNTF Overview DanNotes 11/23/11
OpenNTF Overview DanNotes 11/23/11
 
Annual Report
Annual ReportAnnual Report
Annual Report
 
Data management plans – EUDAT Best practices and case study | www.eudat.eu
Data management plans – EUDAT Best practices and case study | www.eudat.euData management plans – EUDAT Best practices and case study | www.eudat.eu
Data management plans – EUDAT Best practices and case study | www.eudat.eu
 

Último

DevoxxFR 2024 Reproducible Builds with Apache Maven
DevoxxFR 2024 Reproducible Builds with Apache MavenDevoxxFR 2024 Reproducible Builds with Apache Maven
DevoxxFR 2024 Reproducible Builds with Apache MavenHervé Boutemy
 
Passkey Providers and Enabling Portability: FIDO Paris Seminar.pptx
Passkey Providers and Enabling Portability: FIDO Paris Seminar.pptxPasskey Providers and Enabling Portability: FIDO Paris Seminar.pptx
Passkey Providers and Enabling Portability: FIDO Paris Seminar.pptxLoriGlavin3
 
Dev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio WebDev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio WebUiPathCommunity
 
TrustArc Webinar - How to Build Consumer Trust Through Data Privacy
TrustArc Webinar - How to Build Consumer Trust Through Data PrivacyTrustArc Webinar - How to Build Consumer Trust Through Data Privacy
TrustArc Webinar - How to Build Consumer Trust Through Data PrivacyTrustArc
 
Unraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfUnraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfAlex Barbosa Coqueiro
 
Take control of your SAP testing with UiPath Test Suite
Take control of your SAP testing with UiPath Test SuiteTake control of your SAP testing with UiPath Test Suite
Take control of your SAP testing with UiPath Test SuiteDianaGray10
 
A Journey Into the Emotions of Software Developers
A Journey Into the Emotions of Software DevelopersA Journey Into the Emotions of Software Developers
A Journey Into the Emotions of Software DevelopersNicole Novielli
 
"ML in Production",Oleksandr Bagan
"ML in Production",Oleksandr Bagan"ML in Production",Oleksandr Bagan
"ML in Production",Oleksandr BaganFwdays
 
Merck Moving Beyond Passwords: FIDO Paris Seminar.pptx
Merck Moving Beyond Passwords: FIDO Paris Seminar.pptxMerck Moving Beyond Passwords: FIDO Paris Seminar.pptx
Merck Moving Beyond Passwords: FIDO Paris Seminar.pptxLoriGlavin3
 
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)Mark Simos
 
Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Commit University
 
New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024BookNet Canada
 
Moving Beyond Passwords: FIDO Paris Seminar.pdf
Moving Beyond Passwords: FIDO Paris Seminar.pdfMoving Beyond Passwords: FIDO Paris Seminar.pdf
Moving Beyond Passwords: FIDO Paris Seminar.pdfLoriGlavin3
 
"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii Soldatenko"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii SoldatenkoFwdays
 
Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024BookNet Canada
 
Artificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptxArtificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptxhariprasad279825
 
Advanced Computer Architecture – An Introduction
Advanced Computer Architecture – An IntroductionAdvanced Computer Architecture – An Introduction
Advanced Computer Architecture – An IntroductionDilum Bandara
 
How to write a Business Continuity Plan
How to write a Business Continuity PlanHow to write a Business Continuity Plan
How to write a Business Continuity PlanDatabarracks
 
Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024BookNet Canada
 
What is Artificial Intelligence?????????
What is Artificial Intelligence?????????What is Artificial Intelligence?????????
What is Artificial Intelligence?????????blackmambaettijean
 

Último (20)

DevoxxFR 2024 Reproducible Builds with Apache Maven
DevoxxFR 2024 Reproducible Builds with Apache MavenDevoxxFR 2024 Reproducible Builds with Apache Maven
DevoxxFR 2024 Reproducible Builds with Apache Maven
 
Passkey Providers and Enabling Portability: FIDO Paris Seminar.pptx
Passkey Providers and Enabling Portability: FIDO Paris Seminar.pptxPasskey Providers and Enabling Portability: FIDO Paris Seminar.pptx
Passkey Providers and Enabling Portability: FIDO Paris Seminar.pptx
 
Dev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio WebDev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio Web
 
TrustArc Webinar - How to Build Consumer Trust Through Data Privacy
TrustArc Webinar - How to Build Consumer Trust Through Data PrivacyTrustArc Webinar - How to Build Consumer Trust Through Data Privacy
TrustArc Webinar - How to Build Consumer Trust Through Data Privacy
 
Unraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfUnraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdf
 
Take control of your SAP testing with UiPath Test Suite
Take control of your SAP testing with UiPath Test SuiteTake control of your SAP testing with UiPath Test Suite
Take control of your SAP testing with UiPath Test Suite
 
A Journey Into the Emotions of Software Developers
A Journey Into the Emotions of Software DevelopersA Journey Into the Emotions of Software Developers
A Journey Into the Emotions of Software Developers
 
"ML in Production",Oleksandr Bagan
"ML in Production",Oleksandr Bagan"ML in Production",Oleksandr Bagan
"ML in Production",Oleksandr Bagan
 
Merck Moving Beyond Passwords: FIDO Paris Seminar.pptx
Merck Moving Beyond Passwords: FIDO Paris Seminar.pptxMerck Moving Beyond Passwords: FIDO Paris Seminar.pptx
Merck Moving Beyond Passwords: FIDO Paris Seminar.pptx
 
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
 
Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!
 
New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
 
Moving Beyond Passwords: FIDO Paris Seminar.pdf
Moving Beyond Passwords: FIDO Paris Seminar.pdfMoving Beyond Passwords: FIDO Paris Seminar.pdf
Moving Beyond Passwords: FIDO Paris Seminar.pdf
 
"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii Soldatenko"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii Soldatenko
 
Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: Loan Stars - Tech Forum 2024
 
Artificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptxArtificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptx
 
Advanced Computer Architecture – An Introduction
Advanced Computer Architecture – An IntroductionAdvanced Computer Architecture – An Introduction
Advanced Computer Architecture – An Introduction
 
How to write a Business Continuity Plan
How to write a Business Continuity PlanHow to write a Business Continuity Plan
How to write a Business Continuity Plan
 
Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC CataList - Tech Forum 2024
 
What is Artificial Intelligence?????????
What is Artificial Intelligence?????????What is Artificial Intelligence?????????
What is Artificial Intelligence?????????
 

Lime recommendation

  • 1. © Fraunhofer IDMT T. Köllmer†, E. Berndl‡, T. Weißgerber‡, P. Aichroth† , H. Kosch‡ A Workflow for Cross Media Recommendations based on Linked Data Analysis † Fraunhofer IDMT ‡ University of Passau Thomas Köllmer thomas.koellmer@idmt.fraunhofer.de
  • 2. © Fraunhofer IDMT 2 The MICO Project  http://www.mico-project.eu  FP7 STREP, GA# 610480  7 Partners: Salzburg Research, Fraunhofer IDMT, Oxford University, University of Passau, Umeå University, Zaizi, InsideOut10  Duration: 36 Months (end: 10/2016)  Goal of MICO is to develop a platform to analyse “media in context” …  ...by orchestrating different content extraction tools that can work simultaneously or in a sequence and …..  ...provide valuable metadata to third party applications.
  • 3. © Fraunhofer IDMT 3 Cross Media Recommendations in MICO  The Recommendation Problem  “Providing suggestions for items to be of use to a user”  Broad Application domains:  e-Commerce, Travel, Multimedia, News, …  Two main data sources:  User behaviour data (Collaborative Filtering)  Item Metadata (Content based recommendation)
  • 4. © Fraunhofer IDMT 4 Cross Media Recommendations in MICO  Media Recommendation  “Classic” Recommendation problems (Music-Recommendation, Netflix Challenge)  Mostly based on collaborative filtering, but also content analysis  Cross Media Recommendation  Cross Media ≠ Cross Modal  Multiple Input  Media Items + Associated Media context, e.g., commentary  Multiple Output  Not necessarily more of the same thing
  • 5. © Fraunhofer IDMT 5 Editor Support Usecase As an editor, while I create or edit articles using WordPress, I want to automatically get related articles and videos that I might link to the article.  Project Use Case: Recommendations for Greenpeace Magazine Italia  Articles  Videos  User behaviour data
  • 6. © Fraunhofer IDMT 6 Editor Support Usecase As an editor, while I create or edit articles using WordPress, I want to automatically get related articles and videos that I might link to the article.  How can Mico Help?  Articles: NER components for content spotting  Videos  Text to speech + NER  Metadata Analysis  User behaviour data  Generic collaborative filtering component (prediction.io)
  • 7. © Fraunhofer IDMT 7 Chat Analysis Usecase As a zooniverse user, I want get more information that helps me during the current classification task, e.g., by the recommendation of similar subjects or training data.  Input:  Results from animal detection  Sentiment and competence classification of talks  NER for animal spotting  Output:  Classification Hints  Hints for moderators / experts
  • 8. © Fraunhofer IDMT 8 Processing Workflow for Editor Support Use Case 1. Crawler feeds media items to platform e.g., videos from Greenpeace channel, news articles, internal archive…. 2. Analysis results are stored as Linked Data Currently: using Apache Marmotta as triplestore
  • 9. © Fraunhofer IDMT 9 Processing Workflow for Editor Support Use Case 3. Item gets into Focus (someone is writing an article on a specific topic) 4. Focused item is fed to the same platform, using customized pipelines 5. Results are processed by LD matching, i.e Recommendation component.. 6. … enriched by stored annotations … 7. And returned to editing platform
  • 10. © Fraunhofer IDMT 10 Recommendation Workflow in MICO
  • 11. © Fraunhofer IDMT 11 Ontology Matching  Semantic distance of entities important for analysis  Animal detection: “I see a bird on the tree”, “Yes it’s definitely a sparrow”  Topic detection: Wind energy, sustainability, renewables, oil, …  Non-Mico:  Genres! (e.g., currently 1400 genres on Spotify…)  Not less controversial: movie genres
  • 12. © Fraunhofer IDMT 12 Evaluation  How to evaluate a recommender system?  Forget about precision and recall!  Observing user behaviour  Online experiment  Offline: extrapolating past user behavior  Lab Studies / Questionnaire  Open problem: standardized + comparable testing of recommender systems
  • 13. © Fraunhofer IDMT 13 Conclusions  Proposed setting tries to formalize a certain subset of recommendation problems: Cross media recommendation  Focused on Media Items  Profits from Content Analysis  New items, have a high contextual dependency on existing items, that will be exploited for recommendation  Work in Progress:  Stable release of Mico-platform and extractors: right now  PoC recommendation code will be integrated in the platform  Demo Services: available in June  Incorporate feedback into recommendation
  • 14. © Fraunhofer IDMT T. Köllmer†, E. Berndl‡, T. Weißgerber‡, P. Aichroth† , H. Kosch‡ A Workflow for Cross Media Recommendations based on Linked Data Analysis † Fraunhofer IDMT ‡ University of Passau Thank you! Questions?