TextGraphs-6: Graph-based Methods for Natural Language Processing
Workshop at ACL-HLT 2011, Portland, Oregon.
TextGraphs is at its SIXTH edition! This shows that two seemingly distinct disciplines, graph theoretic models and computational linguistics, are in fact intimately connected, with a large variety of Natural Language Processing (NLP) applications adopting efficient and elegant solutions from graph-theoretical framework.
The TextGraphs workshop series addresses a broad spectrum of research areas and brings together specialists working on graph-based models and algorithms for NLP and computational linguistics, as well as on the theoretical foundations of related graph-based methods. This workshop series is aimed at fostering an exchange of ideas by facilitating a discussion about both the techniques and the theoretical justification of the empirical results among the NLP community members.
Special Theme: "Graphs in Structured Input/Output Learning"
Recent work in machine learning has provided interesting approaches to globally represent and process structures, e.g.:
- graphical models, which encode observations, labels and their dependencies as nodes and edges of graphs
- kernel-based machines which can encode graphs with structural kernels in the learning;
algorithms
- SVM-struct and other max margin methods and the structured perceptron that allow
for outputting entire structures like for example graphs
To make such methods effective both from efficiency and accuracy viewpoint, when using
graphs as input and output, the typical graph properties must be exploited at the best.
Our aim is to bring together researcher experts in graph theory and experts in machine
learning for structure predictions from structured input in order to enable cross-fertilization
of ideas. The proposed theme will encourage publication of early results and initiate discussions
of issues in this area. We hope that this will help to shape future directions for
improving both efficiency and accuracy of algorithms dealing with structured input/output,
where graphs are seen as the main tool to model natural language processing data.
Special Issue
High quality papers of Textgraphs-6 that are related to the special theme will be invited to submit to the JNLE special issue on
Statistical Learning of Natural Language Structured Input and Output.
After receiving indications from the workshop organizers the authors will be invited to submit their journal version (see link above for the related guidelines). The schedule would be as indicated below:
- Submission of revised articles: 28 August 2011 (journal version)
- Final decisions to authors: 23 October 2011
- Final versions due from authors: 27 November 2011
Invited Speaker
We are delighted to announce that our invited speaker for TextGraph-6 will be Hal Daumé III.
He will give the talk
Structured Prediction need not be Slow
Suggested topics
We invite submissions on the following (but not limited to) general topics (including those from the special theme):
- Graphical and conditional models, e.g. conditional random field, LDA
- Algorithms for graph output, e.g. structured SVMs, perceptron
- Kernel methods for graphs, e.g. random walk, tree and sequence kernels
- Relational learning using graphs
- Integer linear programming with graph-based constraints
- Automatic analysis of linking structures, e.g. web documents, blogs
- Graph methods for syntactic/semantic parsing
- Learning social graphs
- Graph-based representations, acquisition and evaluation of lexicon and ontology
- Dynamic graph representations for NLP
- Properties of lexical, semantic, syntactic and phonological graphs
- Clustering-based algorithms
- Application of spectral graph theory in NLP
- Unsupervised and semi-supervised learning models based-on graphs
- Graph methods for NLP tasks, e.g. morpho-syntactic annotation, word sense disambiguation,
syntactic/semantic parsing
- Graph methods for NLP applications, e.g. retrieval, extraction and summarization of
information
- Semantic inference using graphs, e.g. question answering and text entailment recognition
Author Instructions
-
Submissions will consist of:
- regular full papers of max. 8 pages (one additional page for the Reference section only is allowed - for a maximum of 9 pages)
- regular short papers of max. 4 pages (one additional page for the Reference section only is allowed - for a maximum of 5 pages)
- position papers of max. 4 pages (describing new scenarios for the use of graphs for text processing, especially in the field of Graphs in Structured Input/Output Learning).
- All submissions must be electronic in PDF and must be formatted using the ACL-HLT 2011 Style Files (see http://www.acl2011.org/call.shtml#submission).
- Multiple-submission policy
- Papers that have been or will be submitted to other meetings or publications must indicate this at submission time.
- Authors submitting multiple papers to TextGraphs-6 may not submit papers that overlap significantly (50%) with each other in content or results.
- Authors of papers accepted for presentation at Textgraphs-6 must notify the organizers immediately as to whether the paper will be presented.
- All accepted papers must be presented at the conference in order to appear in the proceedings.
- Blind review policy
- In order to facilitate blind reviewing, the authors should omit their names and affiliations from the paper. Furthermore, self-references that reveal the author identity, e.g., "We previously showed (Smith, 1991)" must be avoided. Instead, use citations such as "Smith previously showed (Smith, 1991)". Papers that do not conform to these requirements will be rejected without review.
- Submission
Program Committee
| Eneko Agirre | University of the Basque Country, Spain |
| Roberto Basili | University of Rome, Tor Vergata, Italy |
| Ulf Brefeld | Yahoo! Barcelona, Spain |
| Razvan Bunescu | Ohio University |
| Nicola Cancedda | Xerox Research Centre Europe, France |
| Giuseppe Carenini | University of British Columbia, Canada |
| Yejin Choi | Stony Brook University |
| William Cohen | Carnegie Mellon University |
| Andras Csomai | Google USA |
| Mona Diab | Columbia University |
| Gael Dias | Universidade da Beira Interior, Portugal |
| Michael Gamon | Microsoft Research, Redmond |
| Thomas Gaertner | University of Bonn and Fraunhofer IAIS, Germany |
| Andrew Goldberg | University of Wisconsin |
| Richard Johansson | Trento University, Italy |
| Lillian Lee | Cornell University |
| Chris Manning | Stanford University |
| Ryan McDonald | Google Research USA |
| Rada Mihalcea | University of North Texas |
| Animesh Mukherjee | CSL Lab, ISI Foundation, Torino, Italy |
| Bo Pang | Yahoo! Research |
| Patrick Pantel | USC Information Sciences Institute |
| Daniele Pighin | Universidad Politecnica de Catalunya, Spain |
| Uwe Quasthoff | University of Leipzig, Germany |
| Dragomir Radev | University of Michigan |
| Dan Roth | University of Illinois at Urbana Champaign |
| Aitor Soroa | University of the Basque Country, Spain |
| Veselin Stoyanov | Johns Hopkins University |
| Swapna Somasundaran | Siemens Corporate Research, USA |
| Theresa Wilson | University of Edinburgh, UK |
| Min Zhang | Institute for Infocomm Research, A-STAR, Singapore |
Workshop Organizers
- Irina Matveeva, Dieselpoint, Inc, USA
- Email: matveeva {at} cs.uchicago.edu
- Alessandro Moschitti, University of Trento, Italy
- Email: moschitti {at} disi.unitn.it
- Lluís Màrquez, Universitat Politècnica de Catalunya, Spain
- Email: lluism {at} lsi.upc.edu
- Fabio Massimo Zanzotto, University of Rome, Italy
- Email: zanzotto {at} info.uniroma2.it
