CFP: Workshop NeSp-NLP 2010: Negation and Speculation in Natural Language Processing

Workshop NeSp-NLP 2010: Negation and Speculation in Natural Language Processing
July 10, 2010, Uppsala, Sweden

In recent years, research has yielded substantial progress in NLP tasks like NE recognition, WSD, parsing, semantic role labeling, and anaphora resolution among others. This has been in part supported by the organisation of shared tasks, which provide annotated data, a definition of the task and an evaluation framework, motivating researchers to develop new techniques to tackle these tasks. Other tasks like paraphrasing, summarization or textual entailment have also progressed, but results are still relatively low because deep understanding of language - mapping meaning to meaning - is necessary. This raises methodological questions. Furthermore, large scale linguistic resources are still lacking.

Negation and speculation are two phenomena involved in deep understanding of text. Both are related to expressing the factuality of statements, that is, expressing to which extent a statement is or is not a fact or a speculation. Negation turns an affirmative statement into negative (it rains/it does not rain). Speculation is used to express to which extent a statement is certain or speculated (it might rain/apparently, it will rain/ it is likely to rain/it is not clear whether it will rain/we suspect that it will rain).

SCOPE AND TOPICS

In this workshop we aim at bringing together researchers working on negation and speculation from any area related to computational language learning and processing. The general goal of the workshop is to stimulate research about these topics. Specific goals are to describe the lexical aspects of negation and speculation, to define how the semantics of these phenomena can be modelled for computational purposes, to explore techniques aimed at learning the factuality of an statement, and to analyse how the treatment of these phenomena affects the efficiency of NLP applications. Finally, the workshop aims at encouraging discussion about the need of deep linguistic processing as a way to take computational linguistics a step further.

The wokshop will address the following aspects of negation and speculation, although it will be open to other related topics:

• Lexical aspects of negation and speculation
• Linguistic resources with information about negation and speculation: corpora, dictionaries, lexical databases
• Descriptive analysis of negation and speculation cues
• Negation and speculation across domains and genres
• Negation and speculation in biomedical texts and biomedical text mining
• Handling negation and speculation in NLP: dialogue systems, sentiment analysis, text mining, textual entailment, information extraction, machine translation, paraphrasing
• Learning the scope of negation and speculation cues
• Interaction of negation and speculation for evaluating the factuality of an statement
• Corpora annotation: guidelines, bootstrapping techniques, quality assessment
• Modelling factuality for computational purposes
• Algorithms to learn negation and speculation
• Structured prediction of negation and speculation
• Joint learning of negation and speculation
• Inference of factual knowledge

IMPORTANT DATES

• May 14 - Deadline for workshop papers
• June 15 - Notification of acceptance
• June 25 - Camera-ready papers due
• July 10 - Workshop in Uppsala