AI*IA Workshop on "Learning by Reading in the Real World" (LERREW 2011)

co-located with the XII Conference of the Italian Association for Artificial Intelligence (AI*IA 2011 http://chilab.dinfo.unipa.it/aixia2011/)

Palermo, Italy, September 15, 2011

LERREW 2011 Palermo

Currently, almost all successful AI systems need to rely on sufficient relevant knowledge about a problem domain in order to perform a task. Such knowledge needs to be consistent and expressed in a logical formalism of some type. Since manually encoding such knowledge can become prohibitively expensive, the possibility to automatically acquire it from naturally occurring text and to convert it into a deep logical notation would greatly improve the systems’ performance. The NLP community has become increasingly aware of this new framework, also known as “Learning by Reading”, and several events concerning different aspects of this approach have been organized recently, such as the 1st International Workshop on Formalisms and Methodology for Learning by Reading (http://www.isi.edu/~rutu/FAM-LbR.php), the IJCAI-2011 Workshop on Learning by Reading and its Applications in Intelligent Question-Answering (http://www.isi.edu/~rutu/FAM-LbR-KRAQ-2011.php), as well as the CLEF 2011 Evaluation Campaign on Question Answering for Machine Reading Evaluation (http://celct.fbk.eu/QA4MRE/).


With respect to previous events, this workshop will be more focused on the role of background knowledge in Machine Reading applications. More specifically, we particularly encourage the submission of unpublished papers addressing the issues of acquiring and using background knowledge in NLP applications, of exploiting structured data for the semantic annotation of texts and of performing semantic enrichment of documents for deep understanding.

Topics of interest include (but are not limited to) methods and paradigms for:

  • Analysis, merging and linking of existing knowledge sources such as Wikipedia

  • Knowledge extraction from text vs. using pre-built knowledge sources

  • Bridging knowledge gaps in text through inference

  • Identification of implicit knowledge in text including null instantiated semantic roles

  • Bootstrapping learning

  • Portability techniques of machine reading approaches for closed domains

  • Recognizing temporal sequences, casuality and other semantic relations between events in texts

  • Machine-reading based applications, including question answering systems

  • Ontology learning, merging and expansion

  • Intrinsic evaluation of inference methods

The workshop is supported by the interest group on NLP of the AI*IA (https://sites.google.com/a/aixia.it/nlp). Additional sponsorship is foreseen by the LiveMemories project (http://www.livememories.org ).

The workshop organizers are considering the opportunity to publish selected and revised versions of the workshop papers as a special issue of Intelligenza Artificiale - the International Journal of the AI*IA, in the second half of 2012 (http://www.iospress.nl/loadtop/load.php?isbn=17248035 ).


Submission deadline: May 22, 2011

Notification of acceptance: July 19, 2011

Camera-ready papers due: August 7, 2011

Workshop: September 15, 2011


Authors are invited to submit full papers of up to 8 pages (including references) in electronic format. We also solicit short contributions of up to 4 pages presenting challenge or position papers. Papers must be submitted as PDF documents to the following address: https://www.easychair.org/conferences/?conf=lerrew2011. The Latex template can be downloaded here. As reviewing will be blind, please ensure that papers are anonymous.


Professor YORICK WILKS, Florida Institute of Machine and Cognition

"How many classic NLP processes contribute to knowledge by reading?"

Abstract: The paper surveys some aspects of what counts as “learning by reading” and then goes on to describe some detailed work under some of those headings from Sheffield NLP projects with which I have been involved. Several classic NLP tasks can now be reinterpreted as “learning by reading”: information extraction, automatic summarization, question answering, terminology extraction and ontology learning. I will ask whether “learning by reading” is more than a name for the conjunction of these processes.  I will then describe some Sheffield work on ontology learning, real time web access in dialogue processing, question answering and the derivation of much larger language models with skip grams and the possible relationship of those to knowledge extraction.

Organizing Committee

Sara Tonelli – FBK-irst, Trento (contact: satonelli@fbk.eu)

Fabio Zanzotto – University of Rome Tor Vergata

Roberto Basili - University of Rome Tor Vergata

Bernardo Magnini – FBK-irst, Trento

Program Committee

Sara Tonelli – FBK-irst, (contact)

Fabio Zanzotto – University of Rome Tor Vergata

Roberto Basili University of Rome Tor Vergata

Bernardo Magnini – FBK-irst

Steven Bethard, Katholieke Universiteit Leuven

Volha Bryl, FBK-irst

Philipp Cimiano, University of Bielefeld

Claudio Giuliano, FBK-irst

Iryna Gurevych, University of Darmstadt

Rodolfo Delmonte, University of Venice

Leonardo Lesmo, University of Turin

Ruth Mulkar-Mehta, ISI, University of Southern California

Sergei Nirenburg, UMBC 

Marco Pennacchiotti, Yahoo! Inc.

Massimo Poesio, CIMeC

Josef Ruppenhofer, University of Saarbruecken

Gianni Semeraro, University of Bari

Luciano Serafini, FBK-irst

Caroline Sporleder, University of Saarbruecken

Yorick Wilks, Florida Institute of Machine and Cognition