KRAQ 2011 - KRAQ'11: Knowledge and Reasoning for Answering Questions

The workshop will be organized around a few major questions of interest to a number of AI, NLP, speech, dialogue linguistics, HLT and pragmatics people.
List of topics:

This workshop will focus in particular on two major issues in QA in relation with IJCNLP: what are the difficulties raised by Asian languages when used in QA and how to associate QA with dialogue, speech and various non textual modes of communication (e.g. SMS) ?

The following list of topics is indicative:

New types of questions and related Knowledge Representation, pragmatic and linguistic paradigms supported by interactive modalities (speech, SMS, dialogue, etc.), e.g. enquiries about traffic jams, asking for ways to register to an event, etc.
Language processing: question processing, response production, semantic role detection for QA and related marks, reasoning with lexical resources, textual inference, response production (planning and argumentation), language generation (e.g. lexical choice, templates). The specificities of Asian languages must be outlined, for example: lack of interrogative pronoun or mark, lack of precise focus indication, etc.
Speech processing for QA: there is a natural intersection between tasks that can be performed via spoken interaction (e.g. on mobile devices) and tasks involving concise solutions/answers such as QA. These invoke methods for robust query processing in the presence of ASR error rates, concise and effective answer production criteria, QA-oriented dialogue systems and the migration of Spoken Dialogue Systems to the open-domain;
Reasoning aspects: information fusion-integration, models for explanation production and argumentation, levels of knowledge involved (e.g. ontologies, domain knowledge) and multilingual aspects;
Pragmatic dimensions of intelligently answering questions: user intentions, plan and goal recognition and production, conversational implicatures in responses, principles for the design of cooperative systems, especially within interactive situations or with dedicated communication devices.
Media for QA: the same type of information can be conveyed via different media, and the choice of the most suitable medium given an information type (definition, date, procedure) is critical to ensure the relevance and usability of the answer itself.
Multimodality in QA: different modalities (speech, texts, images, etc.) can be used in formulating questions as well as generating answers. This problem raises several questions, among them: how to represent modality interaction, in what situations is multimodality relevant, what kind of modality correspondence should exist between Q and A, etc. Finally, resolving and reasoning with questions containing temporal and spatial information.
User assistance: it is most important to tailor QA systems with user profiles, and to introduce dialogue capabilities when appropriate. The properties and the capabilities offered by communication media for QA should also be investigated in more depth, be it SMS or advanced interfaces. The types and forms responses can have w.r.t. the communication media is an important issue to address.
? Assisted QA: question answering has do be adapted to users with permanent or temporary handicaps. Assisted techniques can help in both sides (producing a question or understanding an answer): helps in formulating a well-formed question, generation of explained answers, etc.
Applications: multimedia question answering, spoken question answering (increasing uncertainty caused by the speech recognition), question answering of semi-structured documents such as Wikipedia, legislation;
Evaluation: end-to-end evaluation of complex questions, intrinsic evaluation of inference methods, data-intensive vs knowledge-intensive methods, portability techniques.
http://www.irit.fr/recherches/ILPL/KRAQ11.htm