CRAC: NAACL Workshop on Computational Models of Reference, Anaphora, and Coreference - June 6th, 2018, New Orleans

Workshop description

Background There has been a lot of research activity in anaphora / coreference resolution in recent years, but once the DAARC series ended, there were no events in Computational Linguistics entirely dedicated to this type of work. The Coreference Beyond Ontonotes (CORBON) workshops held in 2016 (with NAACL) and 2017 (with EACL) partially addressed this need, but their focus was primarily on under-investigated coreference phenomena. The CRAC series of workshops is intended to be an event of interest to the entire anaphora / coreference / reference community.

Objectives The aim of the proposed workshop(s) is to provide a forum where work on all aspects of computational work on anaphora resolution and annotation, including both coreference and types of anaphora such as bridging references resolution and discourse deixis, can be presented. We also intend to attract work on reference (e.g., deictical reference to objects displayed in a multimodal interface).

Important dates

The important dates for CRAC are the same as for every other NAACL 2018 workhop:

• Deadline for submissions: Extended to 9 March 2018 (11:59pm Hawaii Standard Time)
• Notification of acceptance: 2 April 2018
• Camera-ready version due: 16 April 2018
• Workshop date: 6th June 2018

Data and Resources

The ARRAU corpus is distributed by LDC, that will make it available to the participants to this shared task. Participants should register for the shared task by email to Yulia Grishina (grishina@uni-potsdam.de) and Massimo Poesio (m.poesio@qmul.ac.uk). After registering, participants should download the Agreement form from this site, sign it, and send it to LDC, which will then release the data.

Program

• 9:00-9:10 Welcome (Massimo Poesio, Vincent Ng, Maciej Ogrodniczuk)
• 9:10-10:30 Session 1
• 9:10-10:00 Invited Talk: Ana Marasovic. Resolving Abstract Anaphors in Discourse — Uphill Battles with Neural Ranking Models and Automatic Data Extraction

  Abstract: Abstract anaphora resolution (AAR) is a challenging task that aims to resolve anaphoric reference of pronominal and nominal expressions that refer to abstract objects like facts, events, propositions, actions or situations, in the (typically) preceding discourse. A central property of abstract anaphora is that it establishes a relation between the anaphor embedded in the anaphoric sentence and its (typically non-nominal) antecedent. In this talk, I will present a mention-ranking model that learns how abstract anaphors relate to their antecedents with an LSTM-Siamese Net. I will describe how we harvested training data from a parsed corpus using a common syntactic pattern consisting of a verb with an embedded sentential argument. I will show results of the mention-ranking model trained for shell noun resolution and results on an abstract anaphora subset of the ARRAU corpus. The latter corpus presents a greater challenge due to a mixture of nominal and pronominal anaphors and a greater range of confounders. Finally, I will talk about ideas on how the training data extraction method and the mention-ranking model could be further improved for the challenges ahead.

• 10:00-10:30, Berfin Aktaş, Tatjana Scheffler and Manfred Stede. Anaphora Resolution for Twitter Conversations: An Exploratory Study
• 10:30-11:00 Coffee break
• 11:00-12:30 Session 2: Shared Talk, Plural Reference
• 11:00-11:30 Massimo Poesio, Yulia Grishina, Varada Kolhatkar, Nafise Moosavi, Ina Roesiger, Adam Roussel, Fabian Simonjetz, Alexandra Uma, Olga Uryupina, Juntao Yu and Heike Zinsmeister. Anaphora Resolution with the ARRAU Corpus
• 11:30-12:00 Ina Roesiger. Rule and learning-based models for bridging resolution in the ARRAU corpus
• 12:00-12:30 Amir Zeldes. A Predictive Model for Notional Anaphora in English.
• 12:30-14:00 Lunch
• 14:00-15:30 Session 3: Bridging, Discourse deixis, Anaphora in German, Corpus annotation 1.
• 14:00-14:20 Ina Roesiger, Maximilian Köper, Kim Anh Nguyen and Sabine Schulte im Walde. Integrating Predictions from Neural-Network Relation Classifiers into Coreference and Bridging Resolution.
• 14:20-14:50 Janis Pagel and Ina Roesiger. Towards Bridging Resolution in German: Data Analysis and Rule-based Experiments.
• 14:50-15:10 Adam Roussel. Detecting and Resolving Shell Nouns in German.
• 15:10-15:30 Anna Nedoluzhko, Michal Novák and Maciej Ogrodniczuk. PAWS: A Multi-lingual Parallel Treebank with Anaphoric Relations.
• 15:30-16:00 Mid-afternoon snack
• 16:00-17:30 Session 4: Corpus Annotation 2, Cognitive Models
• 16:00-16:30 Michal Novák. A Fine-grained Large-scale Analysis of Coreference Projection.
• 16:30-17:00 Jixing Li, Murielle Fabre, Wen-Ming Luh and John Hale. Modeling Brain Activity Associated with Pronoun Resolution in English and Chinese
• 17:00-17:30 Sharid Loáiciga, Luca Bevacqua, Hannah Rohde and Christian Hardmeier. Event versus entity co-reference: Effects of context and form of referring expression.

Travel awards

A limited number of travel awards are available to student authors of accepted CRAC papers who are affiliated with U.S. institutions.

Enquiries

For all enquiries, please contact m.poesio@qmul.ac.uk

Workshop Organizers

Co-chairs:

• Massimo Poesio (Queen Mary University of London, UK)
• Maciej Ogrodniczuk (Institute of Computer Science, Polish Academy of Sciences, Poland)
• Vincent Ng (University of Texas at Dallas, USA)

Program committee:

• Anders Bjorkelund, University of Stuttgart
• Antonio Branco,  University of Lisbon
• Dan Cristea, A. I. Cuza University of Iasi
• Pascal Denis, MAGNET, INRIA Lille Nord-Europe
• Sobha Lalitha Devi, AU-KBC Research Center, Anna University of Chennai
• Yulia Grishina, University of Potsdam
• Veronique Hoste, Ghent University
• Ryu Iida, National Institute of Information and Communications Technology (NICT), Kyoto
• Varada Kolhatkar, Simon Fraser University
• Katja Markert, Heidelberg University
• Costanza Navaretta, University of Copenhagen
• Anna Nedoluzhko, Charles University in Prague
• Michal Novak,  Charles University in Prague
• Simone Paolo Ponzetto, University of Mannheim
• Sameer Pradhan, cemantix.org and Boulder Learning Inc.
• Marta Recasens, Google Inc.
• Dan Roth, University of Pennsylvania.
• Veselin Stoyanov, Facebook
• Olga Uryupina, University of Trento
• Yannick Versley, IBM
• Sam Wiseman, Harvard University
• Heike Zinsmeister, University of Hamburg

Shared Task

The proposed (series of( workshop(s) will be accompanied, like CORBON, by a shared task. The shared task(s) in 2018 will be on anaphora resolution using ARRAU (Uryupina et al, to appear), a corpus designed to

• cover a more diverse range of genres than typical coreference corpora, including news, task-oriented dialogue, and narrative;
• provide annotations for all noun phrases, including non-referring NPs and singletons (unlike in OntoNotes where only coreferential mentions are an- notated) and carefully annotated their boundaries catering also for non- continuous mentions;
• include annotation of the full range of anaphoric phenomena, including bridging and discourse deixis;
• include the annotation of a variety of morpho-syntactic and semantic attributes of mentions, including genericity.

For more details, see the shared task page .

Shared Task Organizers

• Yulia Grishina, University of Potsdam (chair)
• Varada Kolhatkar, Simon Fraser University
• Anna Nedoluzhko, Charles University in Prague
• Massimo Poesio, Queen Mary University of London
• Adam Roussel, University of the Ruhr at Bochum
• Fabian Simonjetz, University of the Ruhr at Bochum
• Olga Uryupina, University of Trento
• Heike Zinsmeister, University of Hamburg

Previous Events

• CORBON '17
  CORBON 2017 was the 2nd edition of the workshop on Coreference Resolution Beyond OntoNotes, co-located with EACL 2017
• CORBON '16
  CORBON 2017 was the 1st edition of the workshop on Coreference Resolution Beyond OntoNotes, co-located with NAACL 2016

Games4NLP: Using Games and Gamification for Natural Language Processing

Symposium description

Games4NLP is an independent symposium co-located with the 15th European Chapter of the Association for Computational Linguistics (EACL) Conference that will take place on April 3-7, 2017 in Valencia, Spain.

The Games4NLP symposium aims to promote and explore the possibilities for research and practical applications of using games and gamification for Natural Language Processing (NLP). The main objective is to provide a forum for researchers and practitioners to discuss and share ideas regarding how the NLP research community can benefit from using game/gamification strategies. For example, games can be used to collect large numbers of annotations of human language provided that there are sufficient numbers of players who are motivated to play, and these annotations, when aggregated, can be used as labels that replace or compliment the effort of expert annotators. The symposium welcomes the participation of both academics and industry practitioners interested in the use of games and gamification for NLP.

This symposium is aimed at researchers interested in using games and gamification to collect language data and to connect a community who are already developing and maintaining such games.

Attendance to the workshop if free; however, all attendees must register using the Eventbrite page. Registration to EACL is not required.

The potential topics of interest include, but are not limited to:

• Games for collecting data useful for NLP (games in planning, under development, or deployed);
• Presentation and/or analysis of game data from established games;
• Gamification of NLP tasks (Techniques, best practice and evaluation of gamification strategies);
• Player motivation (Strategies for recruitment and retention of players and player profiling);
• Game design (Conceptual design elements and reports of success and/or failure of designs);
• Processing NLP game data (Aggregation methods and strategies for minimising noise and cheating);
• Use of NLP game data (How game-generated data has been used for NLP applications);
• Evaluation of games for NLP (Metrics for evaluating game performance, evaluating player performance, motivation and bias, and evaluating task difficulty);
• Directive strategies that attempt to use game metrics in real-time.

The symposium will feature presentations of works-in-progress and best practice from researchers using games and gamification for Natural Language Processing tasks. Workshop presentations will be grouped into themes such as player motivation, game design, using data generated by games and measuring game performance. Key note speakers to be announced soon.

Location

Games4NLP will be held in the Las Arenas/El Perellonet rooms (2nd floor) of the Sercotel Sorolla Palace Hotel in Valencia, Spain. Map.

Program

1030-1100 Coffee

1100-1110 Welcome - Massimo Poesio

1110-1230 Games under development (Session chair: Jon Chamberlain)
uComp Language Quiz - A Game with a Purpose for Multilingual Language Resource Acquisition (Arno Scharl and Michael Föls)
Testing game mechanics in games with a purpose for NLP applications (Chris Madge, Udo Kruschwitz, Jon Chamberlain, Richard Bartle and Massimo Poesio)
Curating an Open Information Extraction Knowledge Base Using Games with a Purpose (Kevin Forand and Philippe Langlais)

1230-1250 Panel - Challenges of Games for NLP

1250-1400 Lunch and Demos

1400-1520 Player motivation (Session chair: Pontus Stenetorp)
Who wants to play Zombie? A survey of the players on ZOMBILINGO (Karën Fort, Bruno Guillaume and Nicolas Lefebvre)
Why do we Need Games? Analysis of the Participation on a Crowdsourcing Annotation Platform (Alice Millour and Karën Fort)
An Exploratory Study of Data Quality and Participation in a Games-for-Science Game Community (Jason Radford and David Lazer)

1520-1550 Coffee break and Demos

1550-1710 Game evaluation (Session chair: Massimo Poesio)
Making NLP games fun to play using Free to Play mechanics: RoboCorp case study (Dagmara Dziedzic and Wojciech Włodarczyk)
Metrics of games-with-a-purpose for NLP applications (Jon Chamberlain, Richard Bartle, Udo Kruschwitz, Chris Madge and Massimo Poesio)
Towards a Word Sheriff 2.0: Lessons learnt and the road ahead (Galen Han, Mark Menezes, Leah Halaseh, Jędrzej Stuczyński, Daniele Menara, Sebastian Riedel and Pontus Stenetorp)

1710-1730 Panel - Future directions for Games for NLP

1730-1740 Closing comments - Jon Chamberlain

People

Co-chairs:

• Jon Chamberlain (University of Essex, UK)
• Chris Cieri (Linguistic Data Consortium, University of Pennsylvania, US)
• Karën Fort (Université Paris-Sorbonne, France)

Program committee:

• Richard Bartle (University of Essex, UK)
• Johan Bos (University of Groningen, Netherlands)
• Yun-Gyung Cheong (Sungkyunkwan University, South Korea)
• Eric de la Clergerie (INRIA, France)
• James Fiumara (Linguistic Data Consortium, University of Pennsylvania, US)
• Bruno Guillaume (Inria Nancy Grand Est, France)
• Iryna Gurevych (Technische Universität Darmstadt, Germany)
• Ivan Habernal (Technische Universität Darmstadt, Germany)
• Frank Hopfgartner (University of Glasgow, UK)
• Udo Kruschwitz (University of Essex, UK)
• Michael Meder (TU Berlin, Germany)
• Mathieu Lafourcade (LIRMM, France)
• Verena Lyding (EURAC, Italy)
• Lionel Nicolas (EURAC, Italy)
• Massimo Poesio (University of Essex, UK)
• Federico Sangati (Independent researcher, Italy)
• Pontus Stenetorp (University College London, UK)

Submission

The symposium is inviting extended abstract submissions (2 pages maximum, including references) of proposals for talks, demos and posters.

Submissions must be in PDF, and must conform to the official style guidelines for EACL 2017. Submissions do not need to be anonymised. We ask you to use the EACL-2017 LaTeX template or adapt the Overleaf template for those less familiar with LaTex.

Accepted submissions will be made available electronically on this website.

Please email your submission to games4nlp@dali-ambiguity.eu by Monday 27 February 2017.

• **EXTENDED** Deadline for submissions: Monday 27 February 2017, 11:59pm Hawaii Standard Time
• **DELAYED** Notification of acceptance: Weds 8 March 2017
• Final program: 27 March 2017
• Symposium date: 4 April 2017