Endorsed by SIGLEX, SIGSEM             


 Here's the program and the accepted papers (PDF)

The symposium will start on Wednesday 20, November 2013 at 9:15 AM and will end on Friday 22, November 2013 at 1 PM.

The symposium includes keynote talks, regular paper presentations, and a poster session.

We also offer space for and two panels (one on Textual Inference and one on Distributional Semantics) and two tutorials.

You can see the program by clicking  here

Keynote Talks

Eneko Agirre (University of Basque Country)
Text Understanding using Knowledge-Bases and Random Walks

Johan Bos (University of Groningen, The Netherlands)
The Groningen Meaning Bank

Philipp Cimiano (University of Bielefeld, Germany)
Ontology Lexicalization as a core task in a language-enhanced Semantic Web

Peter Clark (Allen Institute for Artificial Intelligence, USA)
From Textual Entailment to Knowledgeable Machines

Ido Dagan (Bar Ilan University, Israel), Bernardo Magnini (FBK, Italy)
Entailment graphs for text exploration

Mona Diab (George Washington University, USA)
Semantic Textual Similarity: past present and future

Patrick Hanks (University of Wolverhampton, UK)
Corpus-driven Lexical Analysis: Norms and Exploitations in Word Use

Elisabetta Jezek (University of Pavia, Italy)
Sweetening Ontologies cont'ed

Mirella Lapata (University of Edinburgh, UK)
Unsupervised Relation Extraction with General Domain Knowledge

Günter Neumann (DFKI, Germany), Sebastian Pado (University of Stuttgart, Germany)
Design and Realization of the EXCITEMENT Open Platform for Textual Entailment

Octavian Popescu (FBK, Italy)
Regular Pattens - Probably Approximately Correct Language Model

Dan Roth (University of Illinois at Urbana-Champaign, USA)
Computational Frameworks for Supporting Textual Inference

Sabine Schulte in Walde (University of Stuttgart, Germany)
Potential and limits of distributional approaches for semantic relatedness


Regular Paper Presentations

Paolo Annesi, Danilo Croce and Roberto Basili.
Towards Compositional Tree Kernels

Elena Cabrio and Serena Villata.
Detecting Bipolar Semantic Relations among Natural Language Arguments with Textual Entailment: a Study.

Tommaso Caselli, Carlo Strapparava, Laure Vieu and Guido Vetere.
Aligning Verb Senses in Two Italian Lexical Semantic Resources

Ekaterina Ovchinnikova, Andrew Gordon and Jerry Hobbs.
Abduction for Discourse Interpretation: A Probabilistic Framework

Nai-Lung Tsao and David Wible.
Word similarity using constructions as contextual features


Poster Session

Amal Alshahrani and Allan Ramsay.
Inference for Natural Language

E. Bastianelli, G. Castellucci, D. Croce, R. Basili.
Textual Inference and Meaning Representation in Human Robot Interaction

Jane Bradbury and Ismail El Maarouf.
An empirical classification of verbs based on Semantic Types: the case of the 'poison' verbs.

Matthew Capetola.
Quantifiers: Experimenting with Higher-Order Meaning in Distributional Semantic Space

A. Dinu and A. Ciobanu.
Alternative measures of word relatedness in distributional semantics

Lorenzo Ferrone and Fabio Massimo Zanzotto.
Linear Compositional Distributional Semantics and Structural Kernels

Florentina Hristea.
On a Dependency-based Semantic Space for Unsupervised Noun Sense Disambiguation with an Underlying Naïve Bayes Model

Ismail El Maarouf and VÌt Baisa.
Automatic classification of semantic patterns from the Pattern Dictionary of English Verbs

Michael Marlen and David Gustafson.
Extending the Semantics in Natural Language Understanding

Márton Miháltz and Bálint Sass.
What Do We Drink? Automatically Extending Hungarian WordNet With Selectional Preference Relations

Vlad Niculae.
Comparison pattern matching and creative simile recognition

V. Niculae and O. Popescu.
Determining is-a relationships for Textual Entailment


  • Tutorial on Textual Entailment (by Sebastian Pado). The first part of the tutorial will cover the basic premises behind, and sketch the current state of, research in textual entailment. The second part of the tutorial will report more specifically on the current work of the EU project EXCITEMENT whose goal is to alleviate the fragmentation of work in the area of textual entailment through the design and implementation of a component-based, multilingual, open-source framework, the EXCITEMENT Open Platform (EOP).
    The tutorial is avalable online: Introduction & Applications; Algorithms; Knowledge; EXCITEMENT Open Platform.
  • Tutorial on Corpus Patterns: Patrick Hanks, Elisabetta Jezek and Octavian Popescu will focus on Corpus Patterns (theoretical and computational aspects of corpus pattern analysis, sense-stable patterns, chain clarifying relationships, frames, argument structures).

Distributional Semantics Panel 
In this panel we focused on two main questions and we search for possible research directions for addressing them.
- What is the relation between classic distributional semantic modelsvand the new generation of "embeddings" constructed by training deep neural networks?
  • Are these qualitatively different models, or just a new implementation of the basic idea? Is there clear evidence that embeddings are better than traditional models in fair comparisons?
  • Are there specific tasks that only these approaches can tackle?
  • What are their levels of explanatory adequacy, if they are taken as models of the way humans use language?
-Are there many types of similarity?
  • Current distributional semantics models are good at giving a global similarity score between words or phrases. However, we know that things can be similar or dissimilar in different respects: "lions" and "stone lions" are similar in shape, often not so similar in size, very different in function; "unicorns missed Noah's boat" and "unicorns didn't miss Noah's boat" are very similar in form (they are 'about' the same situation), very different in consequences. But how should we go about distinguishing different types of similarity?
  • Could there be, for instance, as many similarities as there are 'qualia', in Pustejovsky's sense? Or should we have asmany similarities as there are properties? ("A and B are similar in {size, color, origin, value ...}").
  • How shall we define and extract these different similariety types in a computationally viable way?