JIR-FRRA 2009

Call for Papers Journal of Information Retrieval Special Issue on FOCUSED RETRIEVAL AND RESULT AGGREGATION

GUEST EDITORS Andrew Trotman (andrew@cs.otago.ac.nz) University of Otago, Dunedin, New Zealand

Mounia Lalmas (mounia@acm.org) University of Glasgow, Glasgow, UK

DETAILS Standard document retrieval finds atomic documents, and leaves to the user the task of locating relevant information within a document. Focused retrieval addresses this latter task by providing the user a more direct access to relevant information. Focused retrieval aims to identify not only documents relevant to a user information need, but also where within the document the relevant information is located. It aims to satisfy the user information need and not to just identify documents that satisfy the information need. There are three main forms of focused retrieval: element retrieval, passage retrieval, and question answering. Element Retrieval (also known as XML-IR) can be applied when the documents in the collection contain some kind of markup (such as XML). The retrieval engine will typically exploit the structure to identify the most relevant paragraphs, sections or documents to return as answers to a query. With passage retrieval, the retrieval engine will typically choose the appropriate size of results to return and the location based mostly on the content of the document (and sometimes its structure). Whereas element retrieval and passage retrieval are used for information seeking questions, question answering (QA) aims to answer more fact seeking questions, and makes use of natural language processing techniques.

Question answering has been investigated in TREC, CLEF and NCTIR for many years, and since 2008 in INEX, and is arguably the ultimate goal of semantic web search research for interrogative information needs. Passage retrieval has a long history in information retrieval research, including INEX and the TREC genomics track, but is also important when searching long documents of any kind. Element retrieval has been the core task at INEX, and is now being investigated in the INEX book search track. The aim of element retrieval (XML-IR) is to identify the most relevant document components to return as an answer to the query. It has already been shown that returning several elements together as one answer triggers a stronger user satisfaction than returning a single element on its own. In the relevance in context retrieval task at INEX, the aim is to return documents constructed from their most relevant elements; the elements are aggregated to form one result. More generally, elements or passages from different documents might be selected to form an aggregated result. In Yahoo! Alpha and Google Universal a query already yields results from a variety of different sources including: images, videos, news, and sponsored results; all aggregated into a single results page.

Result aggregation is a form of automatic document construction. Given a set of documents and document components that satisfy a user's information need (perhaps identified using focused retrieval), an aggregator will combine these into a single result. Techniques include multiple-component summarization, meta-search like result presentation, and mixed-media presentation (when searching over heterogeneous collections of, for example, text, images, video, and music). This special issue on Focused Retrieval and Result Aggregation intentionally covers two topics, but in particular we are interested in examining the entire search process from user query through to aggregated document presentation. The aim of this special issue is to present the current state-of-the-art and the most recent developments in focused retrieval, result aggregation, and their relationship. It also aims to offer a thoughtful perspective of the potential and emerging challenges of these two paradigms.

Prospective authors who aim to contribute to this special issue are encouraged to submit original and unpublished papers dealing with Focused Retrieval, Result Aggregation, or the combination of the two. Papers are solicited in any of (but not limited to) the following areas:

etc
 * Algorithmic approaches to focused retrieval and/or result aggregation
 * Relationship between focused retrieval and result aggregation
 * The effect of media, language and context
 * Interface and presentation issues
 * Evaluation, e.g. effectiveness, user-centered
 * Applications, e.g. web search, mobile search, wiki search, wiki linking,
 * Use cases, e.g. education, law, travel, etc

IMPORTANT DATES Papers due: 1 May 2009 Review and revision completed: 1 July 2009 Camera ready paper due: 1 September 2009

SUBMISSION The guidelines for authors and reviewers are available for download from the INRT webpage: http://www.springer.com/10791.

Submissions can be uploaded via http://www.editorialmanager.com/inrt and should be indicated for consideration in the Special Issue on Focused Retrieval and Result Aggregation"

In case you encounter any difficulties while submitting your manuscript online, please get in touch with the responsible Editorial Assistant by clicking on "CONTACT US" in the EditorialManager start page.

Authors are encouraged to contact the guest editors with any questions

This CfP was obtained from WikiCFP