Text Analysis Conference (TAC 2011)

INTRODUCTION

The Text Analysis Conference (TAC) is a series of evaluations and workshops organized to encourage research in Natural Language Processing and related applications, by providing a large test collection, common evaluation procedures, and a forum for organizations to share their results. TAC comprises multiple tracks, each of which focuses on a particular subproblem of NLP. TAC tracks aim to improve end-user tasks, but also include diagnostic and component evaluations situated within the context of end-user tasks.

You are invited to participate in TAC 2011. NIST will provide test data for each track, and track participants will run their NLP systems on the data and return their results to NIST for evaluation. TAC 2011 has three tracks:

1. Knowledge Base Population (KBP)
2. Recognizing Textual Entailment (RTE-7)
3. Summarization

Organizations may choose to participate in any or all of the tracks. The annual conference culminates in a November workshop at NIST in Gaithersburg, Maryland, USA. All results submitted to NIST are archived on the TAC web site, and all evaluations of submitted results are included in the conference proceedings. Dissemination of TAC work and results other than in the conference proceedings is welcomed, but the conditions of participation specifically preclude any advertising claims based on TAC results.

TRACKS and TASKS

TAC 2011 has three tracks:

1. Knowledge Base Population Track (KBP)
   Web page: http://nlp.cs.qc.cuny.edu/kbp/2011/   (archived here)
   Track Coordinators:
   • Heng Ji ([email protected])
   • Ralph Grishman ([email protected])
   
   The goal of the KBP track is to promote research in automated systems that discover information about named entities as found in a large corpus and incorporate this information into a given knowlege base (namely, a KB derived from Wikipedia). The KBP track comprises the following tasks:
   • Entity-Linking Task: Given a name (of a Person, Organization, or Geopolitical Entity) and a document containing that name, determine the KB node for the named entity, adding a new node for the entity if it is not already in the KB. Two variants of the entity-linking task are offered: English-only, and cross-lingual (both English and Chinese documents).
   • Slot-Filling Task: Given a named entity and a pre-defined set of attributes ("slots") for the entity type, augment a KB node for that entity by extracting all new learnable slot values for the entity as found in a large corpus of documents.
   • Temporal Slot-Filling Task: Same as the regular slot-filling task, but also specify time intervals for each extracted slot value. In addition to a full temporal slot-filling task, a diagnostic temporal task is being offered, in which systems are provided with documents and correct slot values and only have to specify the temporal information.
2. Recognizing Textual Entailment Track (RTE-7)
   Web page: http://tac.nist.gov/2011/RTE/
   Track Coordinators:
   • Danilo Giampiccolo ([email protected])
   • Luisa Bentivogli ([email protected])
   
   The goal of the RTE Track is to develop systems that recognize when one piece of text entails another. RTE-7 pursues the direction of recognizing entailment in larger contexts -- a whole document or set of documents. RTE-7 comprises the following tasks:
   • Main and Novelty-Detection Tasks: Determine whether a given sentence -- in the context of an entire document -- entails a given Hypothesis.
   • Knowledge Base Population Validation Task: Determine whether a given document entails a given TAC KBP relation (e.g., "X is married to Y").
3. Summarization Track
   Web page: http://tac.nist.gov/2011/Summarization/
   Track Coordinators:
   • Karolina Owczarzak ([email protected])
   • Hoa Trang Dang ([email protected])
   Multiling Pilot Coordinators:
   • George Giannakopoulos ([email protected])
   • Ilias Zavitsanos ([email protected])
   
   The goal of the Summarization Track is to develop systems that produce coherent summaries of text. The Summarization track comprises the following tasks:
   • Guided Summarization Task: Produce short, coherent summaries of news articles falling into predefined categories, guided by predefined aspects for each category.
   • Automatically Evaluating Summaries of Peers (AESOP) Task: Automatically score a summary for a given metric, including content (Pyramid score), overall responsiveness, and overall readabilty.
   • Multiling Pilot: Develop and apply partially or fully language-independent summarization algorithms to multiple languages, including Arabic, Czech, English, French, Greek, Hindi, and Hebrew.

TRACK REGISTRATION

Organizations wishing to participate in any of the TAC 2011 tasks are invited to register online by June 10, 2011 at:

http://tac.nist.gov/2011/track-app.html

Track participants are advised to register and submit all required agreement forms as soon as possible in order to receive timely access to track resources. Participants should also make sure they are subscribed to the mailing lists for the tracks for which they are registered (see instructions on the individual track home pages). Registration for a track does not commit you to participating in the track, but is helpful to know for planning. Late registration will be permitted only if resources allow.

Any questions about conference participation may be sent to the TAC project manager: [email protected].

WORKSHOP

The TAC 2011 workshop will be held November 14-15, 2011, in Gaithersburg, Maryland, USA. The TAC workshop is a forum both for presentation of results (including failure analyses and system comparisons), and for more lengthy system presentations describing techniques used, experiments run on the data, and other issues of interest to researchers in NLP.

SCHEDULE