Developing Grand Challenges to Advance Crisis Informatics Technologies

NII Shonan Meeting:

@ Shonan Village Center,May 18-21, 2015 CANCELLED

Organizers

  • Fernando Diaz, Microsoft Research, USA
  • Gareth J. F. Jones, Dublin City University, Ireland
  • Patrick P. Meier, Qatar Foundation – QCRI, Qatar

Overview

Description of the meeting

The lives of many people are altered forever-in often tragic and irreversible ways-when they are caught up in an unexpected crisis in their physical environment. More formally, we can define such a crisis event as a temporally acute, unanticipated natural or manmade disaster. In some settings, citizens and governments can seek to prepare for the occurrence of crisis events, for example planning for earthquake response in geologically active areas or the terrorist incidents in politically unstable environments. Such planning can involve response preparation for individual citizens and emergency services. Unfortunately, however detailed such plans are, the exact response required will be unknown and each event involves unique circumstances. Two things that are certain though are that the response needs to be rapid, often very rapid, and that timely collection, collation and distribution of relevant information to all those caught up in a crisis event is key to limiting immediate distress, suffering and loss of life, and potentially the long term effects on individuals and their environment. In broad terms, high levels of situational awareness, for both responders and affected individuals, can prevent a serious crisis developing into a catastrophe.

The increasing instrumentation of the real world with both traditional sensors such as cameras or ‘citizen sensors’ such as social media [4, 7], is creating opportunities for more rapid and granular situational awareness and crisis response [1]. For example, social media has been used disaster statistics [6], including important information about injuries and damage [2, 3]. The effective integration and exploitation of these opportunities to gather and distribute information in crisis situation is captured in the emerging area of crisis informatics. Nevertheless, while realtime data is rapid available and plentiful, it is vulnerable to unreliability [5].

While these preliminary exercises in crisis informatics have demonstrated some advances, relevant information access technologies are not mature enough to be of reliable practical use to the crisis response community. The research community has begun to address these shortcomings via such activities as the “Workshop on Social Web Search and Mining, Analysis under Crisis” held at SIGIR 2011 and the creation of a TREC track focusing on text summarization during crisis events1.

The key technical problem is that all of the core information access technologies, such as information retrieval, multi-source summarization, machine translation, are particularly brittle in the moments immediately after a crisis happens. For example, general use web search engines are designed for persistent queries where accumulative ranking signals such as clickability or PageRank. These engines often cannot react quickly enough to detect, index, and rank fresh documents relevant to a very rapidly developing disaster situation. Information summarization and extraction systems have similarly focused on leveraging massive redundancy in static, reliable content. Unfortunately, information during crisis events is neither static nor reliable.

At the same time, the crisis response community has become increasingly receptive to contributions from the computer science community. Established organizations such as the United Nations2 as well as newer organizations such as the Standby Task Force3 make use of modern technology developed in the computer science community. Conferences like the International Conference on Information Systems for Crisis Response and Management (ISCRAM)4 and and the International Conference of Crisis Mappers (ICCM)5 reflect this direction.

The availability of sensor and information delivery infrastructure, the openness of the crisis response community to information technology, and existing advances in information access technologies of potential value in crisis response, suggests that the time is right for top level researchers and practitioners to develop a research programme for crisis informatics. This will involve exposing researchers in the computer science community to the needs of the crisis response community in order to develop an understanding of what technology is ready to be transferred and what technology requires further research and experimentation.

Relevant advances in hardware and software technologies, as well as information access via algorithms and data resources, are related to a wide range of heterogeneous information sources, e.g. arising not only from from diverse sources such as social media including Twitter and Facebook, and traditional news broadcast and web-based professional media, but also in different media, obviously text, but also audio, spoken, image and video sources, together with sensor data such as location via GPS sensors. Effectively integration of these sources of variable quality, reliability and availability, in a practical setting of potential network failures and overloads presents significant challenges, often considerably beyond those which anticipated in laboratory research and anticipated everyday usage of commercial products and services.

The goal of this Shonen workshop is to bring together top researchers in computer science with high level experts in crisis response to develop a program for advances in Crisis Informatics. We believe that development of this area has the potential to have real impact on the lives of individuals who find themselves caught up in a crisis event.

This workshop has two objectives. First, to develop an understanding of the issues and problems in crisis response which require technical advances in existing information access methods or algorithms, be they incremental or fundamental. Developing a clearly defined set of such problems will allow the computer science research community to focus on advancing state of methods and algorithms in a manner which is specifically aligned with the needs of the crisis response community. Second, we are interested in understanding which problems in crisis response require solutions which already exist, although perhaps only in the research community and explore strategies which could enable them to be rapidly deployed within the crisis response community.

Anticipated outcomes of the meeting include details of unexploited available information access technologies, research objectives for advances in information technologies, suggestions for potential benchmark evaluation tracks of relevance to crisis informatics at activities such as TREC and NTCIR, possible proposals for focused expert research workshops to address specific technical challenges.


1http://www.trec-ts.org

2http://www.unglobalpulse.org

3http://blog.standbytaskforce.com

4http://iscram.org

5http://crisismappers.net


References
[1] N. Ashish, D. V. Kalashnikov, S. Mehrotra, N. Venkatasubramanian, R. Eguchi, R. Hegde, and P. Smyth. Situational awareness technologies for disaster response. In H. Chen, E. Reid, J. Sinai, A. Silke, and B. Ganoz, editors, Terrorism Informatics: Knowledge Management and Data Mining for Homeland Security. Springer, Dec. 2007.

[2] M. Imran, S. Elbassuoni, C. Castillo, F. Diaz, and P. Meier. Extracting information nuggets from disaster-related messages in social media. In 10th International Conference on Information Systems for Crisis Response and Management, 2013.

[3] M. Imran, S. Elbassuoni, C. Castillo, F. Diaz, and P. Meier. Practical extraction of disaster-relevant information from social media. In Proceedings of the 22nd international conference on World Wide Web companion, WWW ’13 Companion, pages 1021-1024, Republic and Canton of Geneva, Switzerland, 2013. International World Wide Web Conferences Steering Committee.

[4] S. B. Liu, L. Palen, J. Sutton, A. L. Hughes, and S. Vieweg. In search of the bigger picture: The emergent role of on-line photo sharing in times of disaster. In Proceedings of the 5th International ISCRAM Conference, 2008.

[5] M. Mendoza, B. Poblete, and C. Castillo. Twitter under crisis: Can we trust what we rt? In Proceeding of the KDD 2010 Workshop on Social Media Analytics, 2010.

[6] T. Sakaki, M. Okazaki, and Y. Matsuo. Earthquake shakes twitter users: real-time event detection by social sensors. In WWW ’10: Proceedings of the 19th international conference on World wide web, pages 851-860, New York, NY, USA, 2010. ACM.

[7] A. Sheth. Citizen sensing, social signals, and enriching human experience. Internet Computing, IEEE, 13(4):87-92, 2009.