Fourth Workshop on
Textual Case-Based Reasoning:
Textual CBR (TCBR) applies the CBR problem-solving methodology to situations where experiences are predominantly captured in text form. Textual case data is prevalent in many practical applications and maintained in a variety of forms including lessons-learnt documents, incident reports, corporate/personal email communications and many unstructured or semi-structured domain-specific text collections. This variety and volume of text content has created the need for an equally diverse range of CBR development tools.
TCBR systems, like conventional CBR systems, must also acquire knowledge sources to enable the key stages of problem-solving (Retrieve, Reuse, Revise and Retain). However, the implementation of the CBR cycle when these knowledge sources consist of unstructured or semi-structured text is particularly challenging. It calls for hybrid approaches that draw from, and extend, research in related areas (e.g. Information Retrieval, Natural Language Processing, Information Extraction, and Text Mining).
The aim of this workshop is to provide a forum for the discussion of trends, research issues and practical experience in TCBR. We are especially interested in issues that go beyond retrieval, such as solution adaptation, explanation, case base maintenance, and other issues. We have made this the workshop "theme". But, in order to encourage discussion of these issues, alongside the invitation to submit "conventional" research and application papers, we also invite papers that address a common problem, which we refer to as the workshop challenge (see workshop challenge).
This workshop is especially intended for researchers in the CBR community, but other related fields, such as Information Retrieval, Natural Language Processing, Information Extraction, Text Mining and the Semantic Web, may also be interested in participating. The growing community of researchers and developers that deal with textual data is the main target for our workshop.
We invite submissions of two types:
We encourage all potential workshop participants, individually or as a research group, to submit a short paper that addresses the workshop challenge (described below). Submission of a research or application paper (above) does not preclude submission of a workshop challenge paper. Accepted challenge papers will appear in the workshop proceedings, along with research and application papers. Short presentations of the accepted challenge papers at the workshop will form the basis for workshop discussions.
The challenge that we propose consists in analysing the corpus of Air Investigation Reports available from the Transportation Safety Board of Canada. The reports are ordered by year and are available in English and French at:
Investigation reports are created by human experts to describe system malfunctions or procedural errors, in this case in Canadian air traffic. They are written in natural language, but may also contain images referenced from the text. They mostly share a common structure; most frequently, they are divided into sections such as Summary, Analysis, Safety Actions, and so on.
Imagine that you are to use this corpus to build a TCBR system
that supports human investigators. A TCBR system might support
the investigators in tasks such as the following:
A non-exhaustive list of the difficulties to be overcome in
building such a TCBR system includes:
You might like to submit, for example, a 2 or 3 page paper that describes some of the problems in more detail; or a 3 or 4 page paper that proposes ways of overcoming some of these problems; or a 4 or 5 page paper that reports actual experience in addressing the problems (perhaps using the jColibri framework). We particularly welcome contributions that address the workshop theme ("Beyond Retrieval").
The workshop organizers will write a short introductory paper for inclusion in the proceedings to explain the workshop challenge. As well as briefly describing the air incident reports, we will cite some of the work that has already been done in this domain.
Papers of both types should be submitted in Springer LNCS format, which is the format required for the final camera ready copy, subject to the following page limits:
Author instructions along with LaTeX and Word macro files are available on the web at http://www.springer.de/comp/lncs/authors.html. Accepted papers will appear in the Workshop Proceedings. Authors should submit a full paper electronically as a PDF file via the online submission facility accessible at http://iccbr07-ws-text.confmaster.net/. The corresponding author has to first register at site. Then s/he can log in and fill in the information regarding the paper. We also plan to edit a special issue of a journal with extended versions of a selection of the accepted contributions.
Papers should be submitted in Springer LNCS format (http://www.springer.de/comp/lncs/authors.html). (Please do not use page numbers.) Please do not exceed the following page limits:
a maximum of 10 pages for research and application papers; a maximum of 5 pages for papers that address the workshop challenge.
By 25th June, authors should submit a zip file via the online submission facility accessible at http://iccbr07-ws-text.confmaster.net/
For LaTeX users, please submit a zip file that contains: