![Bibliography
De Faria Cordeiro, K., Marino, T., Campos, M. L. M., and Borges, M. R. S. (2011) Linked Open Data
and the Design of Information Infrastructure for Emergency Management Systems. ISCRAM 2011,
1-6.
Heim, P. and Thom, D. (2011) SemSor: Combining Social and Semantic Web to Support the Analysis
of Emergency Situations. Proceedings of the 2nd Workshop on Semantic Models for Adaptive
Interactive Systems SEMAIS.
Okolloh, O. (2008) Ushahidi, or ‘testimony’: Web 2.0 tools for crowdsourcing crisis information.
Participatory Learning and Action 59, 65-70.
Ortmann, J., Limbu, M., Wang, D., and Kauppinen, T. (2011) Crowdsourcing Linked Open Data for
Disaster Management. Framework.
http://wiki.crisiscommons.org/wiki/Tweak_the_Tweet [Accessed: 16-Feb-2011]
Crisis Information Management in the Web 3.0 Age 15](https://image.slidesharecdn.com/crisisinformationmanagement-120423171704-phpapp01/85/Crisis-Information-Management-in-the-Web-3-0-Age-15-320.jpg)
![Example from DBPedia showing some
data about Darmstadt [1]
[1] Paulheim & Fürnkranz (2011)](https://image.slidesharecdn.com/crisisinformationmanagement-120423171704-phpapp01/85/Crisis-Information-Management-in-the-Web-3-0-Age-17-320.jpg)
Crisis Information Management in the Web 3.0 Age
- 1. Crisis Information Management in the Web 3.0 Age Axel Schulz Heiko Paulheim Florian Probst April, 23 2012 Telecooperation Lab The work is partly funded by a grant of the German Federal Ministry for Education and Technische Universität Darmstadt Research
- 2. Motivation No Do we need additional rescue squads? Task 1 Task 2 Fragmented Situational Picture Apps Yes ? Additional Information Citizens Web Data Linked Data Experts Social Media • For making informed decisions a decision maker must be aware of all relevant elements in his environment • Currently decision makers are dependent on information provided by onsite rescue squads • The information landscape is changing rapidly, but potentially valuable information sources are not used by the command staff • Heterogeneous and unstructured nature of the data, lack of time to analyze flood of data Crisis Information Management in the Web 3.0 Age 2
- 3. Vision for Decision Making in Emergency Management No Do we need additional rescue squads? Task 1 Task 2 Apps Additional Information Citizens Web Data Yes Linked Data Experts Social Media Enhanced Situational Picture • Combing Social and Semantic Web to provide structured and relevant information for enhancing the situational picture More informed decisions, shorter reaction times, overall increase of successful operations Crisis Information Management in the Web 3.0 Age 3
- 4. Related Work Few approaches for structuring and augmenting information provided by citizens or that utilize data sources as Linked Open Data (LOD) Tweak the Tweet project proposes a syntax to support a more efficient extraction of relevant data Ushahidi as a live example for filtering user-generated information (Okolloh, 2008) Architecture for applying the Linked Data paradigm to represent and process information in emergencies (De Faria Cordeiro, Marino, Campos, Borges, 2011) Texts from social networks are turned into structured information based on Linked Open Data principles and enriched using volunteers (Ortmann, 2011) Social media is constantly crawled and tagged with links to LOD in SemSor (Heim and Thom, 2011), for retrieval of related information Approaches have shortcomings when it comes to deal with the large amount of information from social networks and LOD Crisis Information Management in the Web 3.0 Age 4
- 5. Web 3.0 Process for Crisis Information Management Information Collection Incident Reporter Information Collection Social Media • Gathering information from social platforms • Incident Reporter mobile application for submission of images, I1 audio and textual descriptions as reports from onsite reporters Cannot directly be used to satisfy the information need of the decision maker Unstructured, unsorted information base Information Overflow Crisis Information Management in the Web 3.0 Age 5
- 6. Web 3.0 Process for Crisis Information Management Information Collection Incident Reporter Information Classification Social Media • Reduce the incoming flood of information to a relevant set which suffices the information need I1 • Approach: “Question-guided Relevance Rating” Information • Command staff asks questions to articulate information need, Classification e.g. "Are there still people in the burning building?" • Reports that are helpful for answering questions are considered to be relevant for the command staff I2 Crisis Information Management in the Web 3.0 Age 6
- 7. Web 3.0 Process for Crisis Information Management Information Collection Incident Reporter Pre-Classification Social Media • Pre-filtering of information base by augmenting textual information with semantic annotations and references to I1 entities in Linked Open Data Information • E.g. “It is burning at TUD” -> retrieve the entity for label TUD -> Classification “Technische Universität Darmstadt” Automatic • Result: Similarity score between information objects to pre-sort preclassification I2 Crisis Information Management in the Web 3.0 Age 7
- 8. Web 3.0 Process for Crisis Information Management Information Collection Incident Reporter Crowd-based classification Social Media • Crowdsourcing is applied to filter according to the relevancy for a specific question I1 • Report Classifier application is based on Question-guided Information Relevance Rating Classification • All citizens can access the application from everywhere and rate Automatic a report with respect to usefulness for answering a question preclassification Crowd-based • Relevance model is used to determine a relevance score for classification each rated information object I2 Crisis Information Management in the Web 3.0 Age 8
- 9. Report Classifier Web Application (Demo Session on Tuesday) Crisis Information Management in the Web 3.0 Age 9
- 10. Web 3.0 Process for Crisis Information Management Information Collection Incident Reporter Information Enrichment Social Media • Further augmentation with additional information I1 Explanation Information • Missing Knowledge -> Existing information must not be self- Classification explanatory Automatic preclassification • E.g. additional information like the chemical properties of Crowd-based sulfur for risk assessment can be provided classification • Create non-obstrusive references to entities in Linked Open I2 Datasets Information Enrichment Extension Explanation • Additional background information can be provided Extension • E.g. important buildings near an incident (chemistry plants, hospitals) Crisis Information Management in the Web 3.0 Age 10
- 11. Mashup for Identifying Critical Infrastructure (Demo Session Tuesday) • MICI submitted for the AI Mashup Challenge 2012 (ESWC ‘12) Crisis Information Management in the Web 3.0 Age 11
- 12. Web 3.0 Process for Crisis Information Management Information Collection Incident Reporter Social Media I1 Information Classification Automatic preclassification Result is a structured information base that could Crowd-based classification enhance the situational picture of the decision I2 maker Information Enrichment Explanation Extension I3 Crisis Information Management in the Web 3.0 Age 12
- 13. Conclusion & Outlook • Combination of Social and Semantic Web to provide structured and relevant information for enhancing the situational picture • Reducing the information overflow using crowd-based classification • Linked Open Data for clustering and enhancing information base Future Work User Study: Evaluation of applicability in real environments Apply machine learning methods for pre-classification The information provided by LOD also needs to be classified Alternative ways of clustering of unstructured social sensor streams, e.g. as virtual citizen sensors Crisis Information Management in the Web 3.0 Age 13
- 14. THANK YOU! Questions? Can also be addressed to: [email protected] Crisis Information Management in the Web 3.0 Age 14
- 15. Bibliography De Faria Cordeiro, K., Marino, T., Campos, M. L. M., and Borges, M. R. S. (2011) Linked Open Data and the Design of Information Infrastructure for Emergency Management Systems. ISCRAM 2011, 1-6. Heim, P. and Thom, D. (2011) SemSor: Combining Social and Semantic Web to Support the Analysis of Emergency Situations. Proceedings of the 2nd Workshop on Semantic Models for Adaptive Interactive Systems SEMAIS. Okolloh, O. (2008) Ushahidi, or ‘testimony’: Web 2.0 tools for crowdsourcing crisis information. Participatory Learning and Action 59, 65-70. Ortmann, J., Limbu, M., Wang, D., and Kauppinen, T. (2011) Crowdsourcing Linked Open Data for Disaster Management. Framework. http://wiki.crisiscommons.org/wiki/Tweak_the_Tweet [Accessed: 16-Feb-2011] Crisis Information Management in the Web 3.0 Age 15
- 16. Linked Open Data Principles Use URIs as names for things Use HTTP URIs so that people can look up those names. When someone looks up a URI, provide useful information, using the standards (RDF*, SPARQL) Include links to other URIs. so that they can discover more things. (http://www.w3.org/DesignIssues/LinkedData.html) Crisis Information Management in the Web 3.0 Age 16
- 17. Example from DBPedia showing some data about Darmstadt [1] [1] Paulheim & Fürnkranz (2011)
- 18. Relevance Models Relevance Aware Unguided Question Guided Experts Relevance Rating Relevance Rating Time Rating Rating Cool Down Cool Down Karma Karma Domain Experts