Home

Workshop Extra-Propositional Aspects of Meaning (ExProM) in Computational Linguistics

June 5, 2015. Collocated with NAACL 2015, Denver, CO, USA

http://www.cse.unt.edu/exprom2015/ 

During the last decade, semantic representation of text has focused on extracting propositional meaning, i.e., capturing who does what to whom, when and where. Several corpora are available, and existing tools extract this kind of knowledge, e.g., role labelers trained on PropBank or NomBank. But propositional semantic representations disregard significant meaning encoded in human language. For example, while sentences (1) and (2) below share the same propositional meaning regarding verb carry, they do not convey the same overall meaning. In order to truly capture what these sentences mean, extra-propositional aspects of meaning (ExProM) such as uncertainty, negation and attribution  must be taken into account.

  1. Thomas Eric Duncan likely contracted the disease when he carried a pregnant woman sick with Ebola.
  2. Thomas Eric personally told me that he never carried a pregnant woman with Ebola.

Workshop Description

The Extra-Propositional Aspects of Meaning in Computational Linguistics Workshop focuses on a broad range of semantic phenomena beyond propositional meaning, i.e., beyond linking propositions and their semantic arguments with relations such as agent (who), theme (what) and location (where).

ExProM is pervasive in human language and, while studied from a theoretical perspective, computational models are scarce. Humans use language to describe events that do not correlate with a real situation in the world.
They express desires, intentions and plans, and also discuss events that did not happen or are unlikely to happen. Events are often described hypothetically, and especulation can be used to explain why something is a certain way without a strong commitment. Humans do not always (want to) tell the (whole) truth: they may use deception to hide lies. Devices such as irony and sarcasm are employed to play with words so that what is said is not what is meant. Finally, humans not only describe their personal views or experiences, but also attribute statements to others. These phenomena are not exclusive of opinionated texts. They are ubiquitous in language, including scientific works (Hyland:98) and news as exemplified below:

  1. A better team might have prevented this infection.
  2. Some speculate that this was a failure of the internal communications systems.
  3. Infected people typically don’t become contagious until they develop symptoms.
  4. Medical personnel can be infected if they don’t use protective gear, such as surgical masks and gloves.
  5. You cannot get it from another person until they start showing symptoms of the disease, like fever.
  6. You can only catch Ebola from coming into direct contact with the bodily fluids of someone who has the disease and is showing symptoms.
  7. We’ve never seen a human virus change the way it is transmitted
  8. There is no reason to believe that Ebola virus is any different from any of the viruses that infect humans and have not changed the way that they are spread.

 

Scope and topics

In its 2015 edition, the Extra-Propositional Aspects of Meaning (ExProM) workshop aims at bringing together scientists working on ExProM within computational linguistics. The goal is to attract researchers interested in theoretical frameworks, annotation schemas, modeling and implementing real systems, as well as analyzing the impact of ExProM in natural language processing applications. The workshop also aims at building a bridge between theoretical and computational linguistics. In particular, it will address these topics, although it will also cover related topics:

  •   Knowledge transfer from linguistics to computational linguistics: defining models for ExProM
  •   ExProM and lexical resources
  •   ExProM at the sentence and discourse level: how much context is necessary?
  •   ExProM and subjectivity: how do linguistic resources encode subjectivity?
  •   ExProM across domains: news, scientific texts, legal documents, economy texts
  •   ExProM in user-generated content
  •   ExProM and implicit meaning: what do sentences really mean?
  •   ExProM in spoken language
  •   Negation: verbal/non-verbal, analytic/synthetic, clausal/subclausal and ordinary/metalinguistic; scope and focus
  •   Modality: defining and annotating types for computational linguistics
  •   Factuality: determining factuality changes within and across documents
  •   Veridicity and veridicality: measuring author commitment
  •   Attribution and perspective: determining who says what and their purpose
  •   Irony and sarcasm
  •   (Automatically) extracting ExProM: strategies, resources and difficulties
  •   Supervised, unsupervised and rule-based approaches to extract ExProM
  •   Integrating ExProM in the NLP pipeline
  •   ExProM for NLP applications: does it help?

Organising commitee

– Eduardo Blanco, University of North Texas
– Roser Morante, VU University Amsterdam
– Caroline Sporleder, Universität Trier

Leave a Reply

Your email address will not be published.