SIGGENINLG 2016

The International Natural Language Generation conference (INLG) is the biennial conference of the Special Interest Group on Natural Language Generation (SIGGEN).

The 9th International Natural Language Generation conference will be held in Edinburgh, Scotland, September 5-8, 2016.


New:   Conference Proceedings

New:   Best paper award

New:   Slides of the Deep Learning for NLG tutorial


Generation Challenges: Call for proposals

WebNLG 2016: 2nd International Workshop on Natural Language Generation and the Semantic Web.

CC-NLG: Computational Creativity in Natural Language Generation

Tutorial: Tsung Hsien Wen (University of Cambridge) on Deep Learning and NLG

INLG is organising the first SIGGEN Hackathon

INLG
		2016

EDINBURGH, SEPTEMBER 5-8, 2016


Sponsored by:
ARRIA   WebNLG

Location


Accommodation


There is no official conference hotel for INLG 2016. There are numerous hotels, University-run accommodation options, bed and breakfasts, and hostels near the conference venue (50 George Square) which can be found on all the major booking sites

Hotels

There are many hotels near the conference venue and in central Edinburgh. The following are options within walking distance of the venue at various different prices.

University accommodation

The University runs several hotel / B&B venues using off-season or former student accommodation.

Bed and Breakfast

There are a number of reasonable bed and breakfast options near the conference venue, particularly in the neighborhood of Newington to the southeast. These tend to be competitively priced compared to hotels. Some representative options are listed below.

Hostels

There are a number of hostels catering to backpackers and other visitors. Those near the converence venue tend to be in noisy areas, though.

Call for Participation


Registration fees:
Main conference
regular-early GBP 260
regular-late GBP 350
student-early GBP 180
student-late GBP 260
Full day Hackathon
regular-early GBP 90
regular-late GBP 150
student-early GBP 60
student-late GBP 90
Half Day Workshop (CC-NLG or WebNLG) or tutorial
regular-early GBP 60
regular-late GBP 100
student-early GBP 40
student-late GBP 60

Important dates


INLG 2016 Hackathon


Join us to try your hand at surface realization!

During the workshop+tutorial days at INLG 2016, we will host the first SIGGEN Hackathon, focusing on generation for controlled natural languages (CNLs; Kuhn 2014). Based on a particular natural language, a controlled natural language aims to preserve most of its natural properties while restricting the lexicon, syntax, and/or semantics. These restrictions make hacking together a surface realizer in 24 hours a tractable enough task that we'd like you to try it!

If you're asking yourself what a CNL might look like, you can think of the extended form of Basic English popularized by xkcd's Up-Goer Five comic, which explained the parts of the Saturn V rocket using only the 1000 most common words in the English language. Alternatively, you might consider AirSpeak, the standardized phraseology of air traffic controllers or Attempto Controlled English (ACE), which is a constrained English with a straightforward mapping to first-order logic.

At the hackathon you will split into small teams of 2-5 people and work on each of the elements of a basic surface realizer, starting with basic morphology and progressing through syntax to produce a system which can generate a majority of the forms allowed in the CNL you choose.

We're hoping for a wide range of participants, from new students to senior researchers, to get the community collaborating, allow us to learn from each other, and get our hands dirty developing our own simple realizers.

For more information on controlled natural languages, check out Tobias Kuhn's "A Survey and Classification of Controlled Natural Languages" (arXiv).

Programme


Monday 5 September
Tuesday 6 September
Wednesday 7 September
Thursday 8 September

Tutorial: Deep Learning and NLG [slides]


Tutorial overview:

Neural Network-based approaches or so-called Deep Learning methods have been the major driving force in the recent development of many natural language processing tasks. Modern network architecture designs have conquered the output format limitation of traditional machine learning models and allow us to map from structured raw input features to structured output labels directly. This characteristic is particularly beneficial for NLP tasks that require a sequential or structured output format such as NLG or MT.

The rise of Deep Learning in NLP can be attributed to the success of Seq2Seq model, which can learn to directly map one sequence to the other. Surprisingly, many problems can be modeled by this architecture if one cast them well. What is more interesting is that we can substitute the encoder architecture to model different input formats, such as using a convolutional network as the front end encoder to encode image and generate its caption, or a simple feed-forward network with attention mechanism to model content selection in NLG.

In this tutorial, we are going to focus on applying Deep Learning to NLG. I will first introduce some fundamental network architectures and techniques that are particularly useful for NLG problems. In the second part of it I will talk about a few successful deep NLG models proposed recently in the NLP research community, and draw some connections to models proposed for different tasks. Finally, I will show you some example codes on how to build these models using Theano package.

Outline:

Invited speakers


Best paper award


Best paper:

Towards Generating Colour Terms for Referents in Photographs: Prefer the Expected or the Unexpected?
Sina Zarrieß and David Schlangen

Runners-up:

CALL FOR PAPERS


INLG 2016

THE 9TH INTERNATIONAL NATURAL LANGUAGE GENERATION CONFERENCE

5-8 September 2016, Edinburgh, Scotland, UK

http://www.macs.hw.ac.uk/InteractionLab/INLG2016/

The Special Interest Group on Natural Language Generation is pleased to announce that the 9th International Conference on Natural Language Generation (INLG 2016) will take place in Edinburgh, Scotland from 5-8 September. The first two days (5-6 September 2016) will consist of workshops, tutorials and a possible hackathon, for which separate calls will be issued, and the final two days (7-8 September 2016) will be the main conference.

Invited speakers:
Workshops:
Tutorial
Hackathon:
Special session:

For the main conference, we invite substantial, original, and unpublished submissions on all topics related to natural language generation.

Topics of interest include (but are not limited to):

Requirements

A paper accepted for presentation at INLG 2016 must not have been presented at any other meeting with publicly available proceedings. Submission to other conferences should be clearly indicated on the paper.

Paper Categories

INLG 2016 has three submission categories, long and short papers, and demos:

Submission Instructions

Submissions should follow the two-column format of ACL Anthology proceedings and be anonymised for double blind reviewing. NAACL 2016 offers both LaTeX style files and Microsoft Word templates.

Papers should be submitted electronically through the START conference management system system at: https://www.softconf.com/g/inlg2016

Important Dates

Organizing Committee

Contact

Contact the workshop organizers with questions: inlg2016@googlegroups.com

Accepted papers


Long papers

ABSOLUTE AND RELATIVE PROPERTIES IN GEOGRAPHIC REFERRING EXPRESSIONS
Rodrigo de Oliveira, Somayajulu Sripada and Ehud Reiter

ABSTRACTIVE COMPRESSION OF CAPTIONS WITH ATTENTIVE RECURRENT NEURAL NETWORKS
Sander Wubben, Emiel Krahmer, Antal van den Bosch and Suzan Verberne

AUTOMATIC LABEL GENERATION OF NEWS COMMENT CLUSTERS
Ahmet Aker, Monica Paramita, Emina Kurtic, Adam Funk, Emma Barker, Mark Hepple and Rob Gaizauskas

CROWD-SOURCING NLG DATA: PICTURES ELICIT BETTER DATA
Jekaterina Novikova, Oliver Lemon and Verena Rieser

DISCOURSE-DRIVEN NARRATIVE GENERATION WITH BIPARTITE PLANNING
David Winer and R. Michael Young

ENABLING TEXT READABILITY AWARENESS DURING THE MICRO PLANNING PHASE OF NLG APPLICATIONS
Priscilla Moraes, Kathleen Mccoy and Sandra Carberry

INFUSING NLU INTO AUTOMATIC QUESTION GENERATION
Karen Mazidi

ON THE VERBALIZATION PATTERNS OF PART-WHOLE RELATIONS IN ISIZULU
C. Maria Keet and Langa Khumalo

PARAPHRASE GENERATION FROM LATENT-VARIABLE PCFGS FOR SEMANTIC PARSING
Shashi Narayan, Siva Reddy and Shay B. Cohen

SELECTING DOMAIN-SPECIFIC CONCEPTS FOR QUESTION GENERATION WITH LIGHTLY-SUPERVISED METHODS
Yiping Jin and Phu Le

SIMPLENLG-IT: ADAPTING SIMPLENLG TO ITALIAN
Alessandro Mazzei, Cristina Battaglino and Cristina Bosco

STATISTICAL NATURAL LANGUAGE GENERATION FROM TABULAR NON-TEXTUAL DATASET
joy mahapatra, Sudip Kumar Naskar and Sivaji Bandyopadhyay

SUMMARISING NEWS STORIES FOR CHILDREN
Iain Macdonald and Advaith Siddharthan

THE SENSEI ANNOTATED CORPUS: HUMAN SUMMARIES OF READER COMMENT CONVERSATIONS IN ON-LINE NEWS
Emma Barker, Monica Lestari Paramita, Ahmet Aker, Emina Kurtic, Mark Hepple and Robert Gaizauskas

TOWARDS GENERATING COLOUR TERMS FOR REFERENTS IN PHOTOGRAPHS: PREFER THE EXPECTED OR THE UNEXPECTED?
Sina Zarrieß and David Schlangen

UNSUPERVISED SENTENCE SIMPLIFICATION USING DEEP SEMANTICS
Shashi Narayan and Claire Gardent

WHAT SHOULD BE MENTIONED IN AN IMAGE DESCRIPTION? A LEARNING TO RANK APPROACH TO CONTENT SELECTION
Josiah Wang and Robert Gaizauskas

Short papers

AN ANALYSIS OF THE ABILITY OF STATISTICAL LANGUAGE MODELS TO CAPTURE THE STRUCTURAL PROPERTIES OF LANGUAGE
Aneiss Ghodsi and John DeNero

AUTOMATIC GENERATION OF STUDENT REPORT CARDS
Amy Isard and Jeremy Knox

CATEGORY-DRIVEN CONTENT SELECTION
Rania Mohammed, Laura Perez-Beltrachini and Claire Gardent

CHALLENGES OF ARGUMENT MINING: GENERATING AN ARGUMENT SYNTHESIS BASED ON THE QUALIA STRUCTURE
Patrick Saint-Dizier

COLLECTING RELIABLE HUMAN JUDGEMENTS ON MACHINE-GENERATED LANGUAGE: THE CASE OF THE QGSTEC DATA
Keith Godwin and Paul Piwek

DESIGNING ALGORITHMS FOR REFERRING WITH PROPER NAMES
Kees van Deemter

ENHANCING PTB UNIVERSAL DEPENDENCIES FOR GRAMMAR-BASED SURFACE REALIZATION
David L. King and Michael White

EVALUATIVE PATTERNS EXTRACTION FOR AUTOMATED TEXT GENERATION
Chia-Chen Lee and Shu-Kai HSIEH

EXPLORING DIFFERENT PREPOSITION SETS, MODELS AND FEATURE SETS IN AUTOMATIC GENERATION OF SPATIAL DESCRIPTIONS
Anja Belz, Adrian Muscat and Brandon Birmingham

GENERATING ENGLISH FROM ABSTRACT MEANING REPRESENTATIONS
Nima Pourdamghani, Kevin Knight and Ulf Hermjakob

GENERATING SUMMARIES OF HOSPITALIZATIONS: A NEW METRIC TO ASSESS THE COMPLEXITY OF MEDICAL TERMS AND THEIR DEFINITIONS
Sabita Acharya, Barbara Di Eugenio, Andrew D Boyd, Karen Dunn Lopez, Richard Cameron and Gail M Keenan

GOOD AUTOMATIC AUTHENTICATION QUESTION GENERATION
Simon Woo

IMPROVING FLUENCY IN NARRATIVE TEXT GENERATION WITH GRAMMATICAL TRANSFORMATIONS AND PROBABILISTIC PARSING
Emily Ahn, Fabrizio Morbini and Andrew Gordon

RANKING AUTOMATICALLY GENERATED QUESTIONS USING COMMON HUMAN QUERIES
Yllias Chali and Sina Golestanirad

STATISTICS-BASED LEXICAL CHOICE FOR NLG FROM QUANTITATIVE INFORMATION
Xiao Li, Kees van Deemter and Chenghua Lin

TASK DEMANDS AND INDIVIDUAL VARIATION IN REFERRING EXPRESSIONS
Adriana Baltaretu and Thiago Castro Ferreira

TENSE AND ASPECT IN RUNYANKORE USING A CONTEXT-FREE GRAMMAR
Joan Byamugisha, C. Maria Keet and Brian DeRenzi

THE MULTILINGUAL AFFECTIVE SOCCER CORPUS (MASC): COMPILING A BIASED PARALLEL CORPUS ON SOCCER REPORTAGE IN ENGLISH, GERMAN AND DUTCH
Nadine Braun, Martijn Goudbeek and Emiel Krahmer

TOWARDS PROPER NAME GENERATION: A CORPUS ANALYSIS
Thiago Castro Ferreira, Sander Wubben and Emiel Krahmer

WHEN TO PLUMMET AND WHEN TO SOAR: CORPUS BASED VERB SELECTION FOR NATURAL LANGUAGE GENERATION
Charese Smiley, Vassilis Plachouras, Frank Schilder, Hiroko Bretz, Jochen Leidner and Dezhao Song

Demos

AUTOMATIC REPORTS FROM SPREADSHEETS: DATA ANALYSIS FOR THE REST OF US
Pablo Duboue

QGASP: A FRAMEWORK FOR QUESTION GENERATION BASED ON DIFFERENT LEVELS OF LINGUISTIC INFORMATION
Hugo Patinho Rodrigues, Luisa Coheur and Eric Nyberg

INCREMENTAL GENERATION OF VISUALLY GROUNDED LANGUAGE IN SITUATED DIALOGUE (DEMONSTRATION SYSTEM)
Yanchao Yu, Arash Eshghi and Oliver Lemon

GENERATION CHALLENGES: CALL FOR PROPOSALS


GenChal Special Session at INLG 2016

5-8 September 2016, Edinburgh, Scotland, UK

http://www.macs.hw.ac.uk/InteractionLab/INLG2016/

Background

Shared tasks have been organised within the Natural Language Generation community since 2007, and have typically been presented at special sessions held during one of the NLG annual events. To date, shared tasks have focussed on a variety of sub-tasks of NLG, including, among others:

GenChal @ INLG 2016

The INLG 2016 organisers and the ACL Special Interest Group in NLG (SIGGEN), are soliciting proposals for new shared tasks. These tasks are expected to run during 2016-17, with a view to presenting results in 2017 during a specially organised GenChal session during one of the NLG annual conferences or workshops.

Accepted proposals will be presented during a special session dedicated to GenChal at INLG 2016.

Submissions

Proposals are expected to describe a shared task in any area that is deemed relevant to NLG. Proposals need not be anonymous, but acceptance will be contingent on an evaluation by a small panel of reviewers.

Proposals should contain the following information:

  1. The names and affiliations of the organisers;
  2. A description of the task, with particular reference to its relevance for the NLG community;
  3. A description of the data that will be provided for participants;
  4. A description of the evaluation methods that will be used to compare peer systems.

GenChal proposals are expected to be no longer than 3 pages and should follow the formatting guidelines for the INLG 2016 conference.

Important dates

CALL FOR WORKSHOP PROPOSALS


INLG 2016

THE 9TH INTERNATIONAL NATURAL LANGUAGE GENERATION CONFERENCE

5-8 September 2016, Edinburgh, Scotland, UK

http://www.macs.hw.ac.uk/InteractionLab/INLG2016/

For the first time since its inception, the International Conference on Natural Language Generation (INLG) will this year include independently organised workshops. These will take place from 5-6 September, immediately prior to the conference itself. In view of this happening for the first time, one workshop, the 2nd International Workshop on NLG and the Semantic Web, has already been confirmed, following an invitation to the organisers by the SIGGEN committee.

The INLG organisers and SIGGEN invite further proposals for one-day or half-day workshops at INLG 2016, which will be held in Edinburgh. Workshops can focus on any topic that is viewed as relevant to the Natural Language Generation community. We especially invite proposals for workshops in relatively new areas which have recently begun to attract interest in the research community, or topics where greater interaction between the NLG community and another established research community can be beneficial to both fields. Proposals for workshops which form part of an existing series are also welcome.

Proposal submission

Proposals should be sent directly to the INLG 2016 Workshop Chair on wp.inlg2016@gmail.com

Workshops at INLG 2016 should be planned as half-day or full one-day events and can take place on September 5th or 6th. Proposals should be approximately two pages long, and should include the following information:

Further details on INLG 2016 are available on the conference website: http://www.macs.hw.ac.uk/InteractionLab/INLG2016/

Important dates

*Suggested* timeline for workshop organisers following acceptance:

Instructions for authors

Submissions should follow the two-column format of ACL Anthology proceedings and be anonymised for double blind reviewing. NAACL 2016 offers both LaTeX style files and Microsoft Word templates

Papers should be submitted electronically through the START conference management system at: https://www.softconf.com/g/inlg2016

Poster Printing

The Teviot Print shop is a few minutes from the conference venue and prints posters from a pdf for anyone, and the prices are:

They also print on an acrylic paper which can be folded for £30.

They are open 9-5 on weekdays and 10-3 on Saturdays, and can normally print a poster in 15 minutes.

http://www.teviotprintshop.com/map.html