NLP architectures in the age of end-to-end deep-learning systems

NLP-ARCHES 2020


Computational Linguistics Data Mining & Analysis Engineering & Computer Science (General) Software Systems



In the past, almost all NLP tasks were tackled with a pipeline approach combining within each pipeline different systems solving clearly defined tasks (e.g. POS tagging, parsing, semantic role labeling etc.). As the different implementations often were not directly compatible several NLP frameworks were introduced (e.g. GATE, DKPro Core, NLTK, spaCy, etc.) that provided the “glue” to hold everything together.

Recently, the trend has clearly focused on end-to-end deep learning systems that directly solve the task at hand without explicitly representing intermediate, linguistically motivated levels. Of course this view is an exaggeration, as the two worlds are still interconnected. Recently, on the one hand, NLP pipeline elements have been discovered in end-to-end systems, on the other hand, end-to-end systems have been integrated in traditional pipeline architectures.

The workshop is motivated by the perceived growing divide between classical pipeline architectures and the recent focus on end-to-end approaches. We believe that there is more common ground than currently considered on both sides. It would be beneficial to bring all interested researchers together at this workshop to explore future directions.


Topics
------
NLP pipeline architectures
End-to-end architectures
Limitations of either approach
Reproducibility issues arising from architecture choices
Orchestration of NLP solutions
Scalability issues
Environmental impact
Explainability issues
Flexibility / extensibility
Licensing / integration issues
Sensitivity to adversarial attacks
Required amount of training data
(In)ability to update models / continue training
Cross programming language issues
Interoperability and compositionality of components


Submission
----------
Participants are asked to submit their work in short paper format. All individual submissions will be presented as posters. Participants will be asked to specifically address the questions from the call (i.e. relationship to frameworks and end-to-end systems etc).

Papers should be formatted according to the stylesheet to be provided at the author kit section on the LREC 2020 website and should not exceed 8 pages, including references and appendices. While we do not have a dedicated track for short papers, we also welcome papers making a contribution to the topics of interest in less than 8 pages. We also welcome system descriptions, opinion pieces, or surveys. If in doubt whether your paper is relevant for the workshop, please get in touch.

Papers should be submitted in PDF format through the START system:
https://www.softconf.com/lrec2020/NLP-ARCHES/

As usual with LREC, reviewing is single-blind, i.e. authors are not anonymous, but reviewers will be. Especially for describing ongoing work on frameworks or tools it is often highly impractical to hide authorship.

Accepted papers will be presented as posters (dimensions TBA). All accepted papers will be published in the workshop proceedings.


"Identify, Describe and Share your LRs!" initiative
----------------------------------------------------
Describing your Language Resources (LRs) in the LRE Map is now a normal practice in the submission procedure of LREC (introduced in 2010 and adopted by other conferences). To continue the efforts initiated at LREC 2014 about “Sharing LRs” (data, tools, web-services, etc.), when submitting a paper, authors will have the possibility to upload LRs in a special LREC repository. This effort of sharing LRs, linked to the LRE Map for their description, may become a new “regular” feature for conferences in our field, thus contributing to creating a common repository where everyone can deposit and share data.

As scientific work requires accurate citations of referenced work so as to allow the community to understand the whole context and also replicate the experiments conducted by other researchers, LREC 2020 endorses the need to uniquely Identify LRs through the use of the International Standard Language Resource Number (ISLRN, www.islrn.org), a Persistent Unique Identifier to be assigned to each Language Resource. The assignment of ISLRNs to LRs cited in LREC papers will be offered at submission time.

Important dates
---------------
Submission deadline : February 14, 2020
Author notification: March 13, 2020
Camera ready: April 2, 2020
Workshop: May 12, 2020


Program Committee
-----------------
Adrien Barbaresi, Berlin-Brandenburg Academy of Sciences
Annemarie Friedrich, Bosch GmbH
Bernhard Schröder, University of Duisburg-Essen
Bryan Jurish, Berlin-Brandenburgische Akademie der Wissenschaften
Chris Hokamp, AYLIEN
Christian Federmann, Microsoft Research
Daniel Baer, Continental
Diana Maynard, University of Sheffield
Dimitris Galanis, ILSP, Athens
Elena Lloret, University of Alicante
Fred Blain, University of Sheffield
Jens Grivolla, Universitat Pompeu Fabra
Jin-Dong Kim, Research Organization of Information and Systems
Johannes Hoffart, Goldman Sachs
John McCrae, National University of Ireland Galway
Kristina Kocijan, University of Zagreb
Maarten van Gompel, Radboud University
Marcos Zampieri, Rochester Institute of Technology
Mohamed Ali HADJ TAIEB, Sfax University
Nancy Ide, Vassar College
Nicolai Erbs, Lead Consultant, INFOMOTION
Nicoletta Calzolari, ILC-CNR
Peter Klügl, Averbis
Renaud Richardet, Frontiers Media
Robert Bossy, INRA
Roman Schneider, IDS Mannheim
Simon Mille, Pompeu Fabra University
Steven Bethard, University of Arizona
Tommaso Teofili, Adobe
Vlad Niculae, Instituto de Telecomunicacoes, Lisbon
Yohei Murakami, Ritsumeikan University


Organizing Committee
--------------------
Piush Aggarwal, Language Technology Lab, University of Duisburg-Essen, Germany
Mohamed Karim Bouzoubaa, ALELM Lab, University Mohammed V in Rabat, Morocco
Richard Eckart de Castilho, UKP Lab, Technische Universität Darmstadt, Germany
Torsten Zesch, Language Technology Lab, University of Duisburg-Essen, Germany


Contact Person
--------------
torsten.zesch@uni-due.de