The second KYOTO workshop Advanced information systems for sharing information and knowledge about the environment

Event Notification Type: 
Call for Abstracts
Abbreviated Title: 
Second Kyoto Workshop
Location: 
Nagaragawa Convention Center
Tuesday, 25 January 2011 to Friday, 28 January 2011
Country: 
Japan
Contact Email: 
City: 
Gifu
Contact: 
Piek Vossen
Submission Deadline: 
Wednesday, 15 December 2010

The second KYOTO workshop
Advanced information systems
for
sharing information and knowledge about the environment
Call for abstracts

Date: 25th - 28th of January 2010
Venue: Nagaragawa Convention Center (http://www.g-ncc.jp/e-framepage.htm), Gifu, Japan (http://en.wikipedia.org/wiki/Gifu,_Gifu).
Travel: Gifu is about one hour from Chubu Central Airport, which has direct flight connections to major cities incl. Frankfurt.
Hosted and sponsored by Toyohashi University of Technology (TUT),
Japan(http://www.tut.ac.jp/english/) and the Asian-European project KYOTO
(http://www.kyoto-project.eu/).

This workshop is subsidized by The National Institute of Information and
Communications Technology (NICT), Japan (http://www.nict.go.jp/index.html).

The goal of the workshop is to demonstrate the possibilities of advanced information systems for sharing information and knowledge about the environment on a global scale. The environment is a global concern crossing boundaries, cultures and languages. Whereas each regional setting is unique, there are many benefits from sharing knowledge and experiences across these regions. This workshop offers a showcase for emerging systems that can enable environmentalist to access this knowledge and information in novel ways. It hopes to bring together technology providers and environmental specialists to discuss the opportunities in the field.

The workshop is organized around the topic of estuaries in the world. A selection of estuaries across the world is presented and innovative applications are shown that can process and present knowledge on these estuaries in useful ways. Participants and guest speakers are invited to use a given set of English text documents about two estuaries: the Humber Estuary in the UK and the Chesapeake Bay in the US. Any analysis of these documents that generates useful knowledge in these documents is welcomed.

The challenge is to give a show-case demonstration of extracted environmental events involving estuaries and present this to the users in a novel and useful way. An event is specified in terms of where, when, what happened and how can these events be grouped. What can refer to entities such as: pollution, human development, human activities, alien species, sedimentation, erosion, increase and decrease of species populations, degradation of habitats.

We provide you with sources of documents on a selection of estuaries but you are free to use your own data sources. There are also no restrictions on what you do with the data as long as it includes text mining and provides output that non-technical users (but perhaps specialists in the domain) can handle and use for their daily work. You should typically focus on more complex and less basic facts than is usually the case in text mining.

Call for abstracts
We invite abstract submissions on the following topics:

Text mining of events and facts in the domain of the environment.
Concept acquisition and knowledge integration in the domain of the environment.
User interfaces for complex knowledge bases with factual information on the environment related to time and place.
Semantic search or question answering systems that allow users to find answers in databases with environmental data
Community building or collaboration environments for building knowledge bases in the domain of the environment.

Abstracts should be 800-1000 words and in PDF format only. Abstracts can be mailed to: p.vossen [at] let.vu.nl. Evaluation is based on relevance to the topics and the estuary-case and on scientific quality.

Event extraction task
In parallel to the general papers of the workshop, we will also organize a specific event extraction task in which you can participate. Participants will be asked to extract events from a selection of textual sources. Events should be represented in the form of simple triplets, which consists of:

a relation
a list of word token identifiers that express the event
a list of word token identifiers that express a participant

Word tokens identifiers are taken from the representation of the text documents in the KYOTO annotation format (see www.kyoto-project.eu and the documentation below for more details). Typically, a single event has multiple participants. Each participant relation is converted to a separate triplet. An example file with annotations and triplets is provided in the task-data. Participants of the workshop can participate in this task by processing the test documents (represented in the KyotoAnnotationFormat) and converting the output into the triplet format. The results should be mailed to: p.vossen [at] let.vu.nl

We already provide now the full English estuary database in the KYOTO annotation format. An example triplet file with annotated events, the annotation tag set and the sources for the estuaries can be downloaded from:

http://kyoto.let.vu.nl/~kyoto/files/kyoto-text-mining-task/Kyoto-text-mi...
(63.9 MB download)

Note that the documents contain a lot of different types of information, both related to the social and political debate about the environment and the organization and related to the physical changes and states of the environment. The workshop focuses on the latter and the test annotation is also restricted to this. The example annotation shows that documents may contain little information about the physical world compared to the information related to the social and political debate. Only information related to the physical world will be considered.

Time schedule
The schedule for the task and abstract submission is as follows:

1. October 1st: release of documents and the example data
2. November 23rd: release of the test files
3. December 8th: deadline for submitting results
4. December 15th: deadline for submission of abstracts
5. December 22nd: results are made public
6. December 22nd: notification of acceptance of abstracts

We will invite a selection of participants to contribute a paper based on their presentation for publication in a book or as a special issue in a journal.

Global program of the workshop:
Before the workshop there will be an excursion to Shirakawa and Takayama so that we get to know each other (January 25). The workshop itself spans three days. For Day 1 (January 26), we will invite major players in information and knowledge disclosure to present their solutions for the environment case. How can they help the environment through information management? We will have discussion panels about the presented cases to collect direct feedback from technology developers and from environmentalists as potential users.

The second day of the workshop (January 27) will be devoted to a more technical discussion on the different aspects of the information and knowledge management. It is divided into text mining, concept mining, standards and wiki-systems and their potential for the future.

The third day (January 28) will focus on local research in Japan related to knowledge mining and the environment.

Program Committee
Piek Vossen, VU University Amsterdam, Netherlands
Histoshi Isahara, Toyohashi University of Technology, Japan
Chu-Ren Huang, Hong Kong Polytechnica University, China
Christiane Fellbaum, Princeton/Berlin Brandenburg Academy of Science
Eneko Agirre, EHU, Spain
German Rigau, EHU, Spain
Monica Monachini, CNR-ILC, Italy
Andrea Marchetti CNR-IIT, Italy
Carlo Aliprandi, Synthema, Italy
Lawrence Jones-Walters, ECNC, Netherlands

Further information
A first demonstrator for a semantic search application can be seen here to exemplify the kind of extraction we have in mind:

http://kyoto.irion.nl/kyoto-IIII/

The following fragments of text taken from a document on the Chesapeake Bay show the type of information that we need to be able to extract:

The Bay’s bottom is home to many species including worms, small fish and shellfish such as clams, oysters and mussels. These bottom-dwelling creatures are especially sensitive to increased pollution and decreased oxygen. These species serve as food for bottom-feeding fish and crabs. The health of these creatures is a good indicator of long-term conditions in the bottom habitat and the Bay overall, because they do not move great distances and have certain predictable responses to environmental stresses.

From such a fragment, we learn facts such as:

worms, small fish, shellfish and clams live on the bottom of the Bay
worms, small fish, shellfish and clams are sensitive to increase of pollution
worms, small fish, shellfish and clams are sensitive to decrease of oxygen
bottom-feeding fish and crabs eat worms, small fish, shellfish and clams
worms, small fish, shellfish and clams do not move great distances

Underwater bay grasses serve many essential ecological functions and are among the most closely monitored habitats in the Bay. Grasses provide critical shelter to many key species such as young striped bass and blue crabs, improve water clarity by helping sediment settle to the bottom, add oxygen to the water and reduce shoreline erosion. Bay grass abundance is an excellent barometer of the health of the Bay because these grasses depend on good local water quality and provide significant benefits to aquatic life.

From this fragment we can learn facts such as:

underwater grass shelters key species such as young striped bass and blue crabs,
underwater grass improves water clarity by helping sediment settle to the bottom
underwater grass adds oxygen to the water
underwater grass reduces shoreline erosion

Other information we learn is where and when these events take place. This is represented through named-entity recognition of dates and locations in the surroundings of the expression of events. For example, Bay here refers to the Chesapeake Bay in the US.