1st Workshop on Language Technologies for Historical and Ancient LAnguages (LT4HALA)
LT4HALA is a one-day workshop that seeks to bring together scholars who are developing and/or are using Language Technologies (LTs) for historically attested languages, so to foster cross-fertilization between the Computational Linguistics community and the areas in the Humanities dealing with historical linguistic data, e.g. historians, philologists, linguists, archaeologists and literary scholars. Despite the current availability of large collections of digitized texts written in historical languages, such interdisciplinary collaboration is still hampered by the limited availability of annotated linguistic resources for most of the historical languages. Creating such resources is a challenge and an obligation for LTs, both to support historical linguistic research with the most updated technologies and to preserve those precious linguistic data that survived from past times.
Relevant topics for the workshop include, but are not limited to:
- handling spelling variation;
- detection and correction of OCR errors;
- creation and annotation of digital resources;
- morphological/syntactic/semantic analysis of textual data;
- adaptation of tools to address diachronic/diatopic/diastratic variation in texts;
- teaching ancient languages with NLP tools;
- NLP-driven theoretical studies in historical linguistics;
- evaluation of NLP tools.
Just because of the limited amount of data preserved for historical and ancient languages, an important role is played by evaluation practices, to understand the level of accuracy of the NLP tools used to build and analyze resources. Given the prominence of Latin, by virtue of its wide diachronic and diatopic span covering two millennia all over Europe, the workshop will host the first edition of EvaLatin, an evaluation campaign entirely devoted to the evaluation of NLP tools for Latin. The first edition of EvaLatin will focus on two tasks (i.e. Lemmatization and PoS tagging), each featuring three sub-tasks (i.e. Classical, Cross-Genre, Cross-Time). These sub-tasks are designed to measure the impact of genre and diachrony on NLP tools performances, a relevant aspect to keep in mind when dealing with the diachronic and diatopic diversity of Latin. Participants will be provided with shared data in the CoNLL-U format and the evaluation script.
For the workshop, we invite papers of different types such as experimental papers, reproduction papers, resource papers, position papers, survey papers. Both long and short papers describing original and unpublished work are welcome. Long papers should deal with substantial completed research and/or report on the development of new methodologies. They may consist of up to 8 pages of content plus 2 pages of references. Short papers are instead appropriate for reporting on works in progress or for describing a singular tool or project. They may consist of up to 4 pages of content plus 2 pages of references. We encourage the authors of papers reporting experimental results to make their results reproducible and the entire process of analysis replicable, by making the data and the tools they used available. The form of the presentation may be oral or poster, whereas in the proceedings there is no difference between the accepted papers. The submission is NOT anonymous. The REC official format is requested. Each paper will be reviewed but three independent reviewers.
As for EvaLatin, participants will be required to submit a technical report for each task (with all the related sub-tasks) they took part in. Technical reports will be included in the proceedings as short papers: the maximum length is 4 pages (excluding references) and they should follow the LREC official format. Reports will receive a light review (we will check for the correctness of the format, the exactness of results and ranking, and overall exposition). All participants will have the possibility to present their results at the workshop: we will allocate an oral session and a poster session fully devoted to the shared tasks.
- 17 February 2020: submissions due
- 10 March 2020: notifications to authors
- 27 March 2020: camera-ready due
- 12 May 2020: workshop
- 10 December 2019: training data available
- Evaluation Window I – Task: Lemmatization
- 17 February 2020: test data available
- 21 February 2020 system results due to organizers
- Evaluation Window II – Task: PoS tagging
- 24 February 2020: test data available
- 28 February 2020: system results due to organizers
- 6 March 2020: assessment returned to participants
- 27 March 2020: reports due to organizers
- 10 April 2020: camera ready version of reports due to organizers
- 12 May 2020: workshop
Share your LRs!
Describing your 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.), authors will have the possibility, when submitting a paper, 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), 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.
- Marcel Bollmann, University of Copenhagen; Denmark;
- Gerlof Bouma, University of Gothenburg, Sweden;
- Patrick Burns, University of Texas at Austin, USA;
- Oksana Dereza, Insight Centre for Data Analytics, Ireland;
- Stefanie Dipper, Ruhr-Universität Bochum, Germany;
- Hanne Eckoff, Oxford University,UK;
- Maud Ehrmann, EPFL, Switzerland;
- Hannes A. Fellner, Universität Wien, Austria;
- Heidi Jauhiainen, University of Helsinki, Finland;
- Julia Krasselt, Zurich University of Applied Sciences, Switzerland;
- John Lee, City University of Hong Kong;
- Chao-Lin Liu, National Chengchi University, Taiwan;
- Barbara McGillivray, University of Cambridge, UK;
- Beáta Megyesi, Uppsala University, Sweden;
- So Miyagawa, University of Göttingen; Germany;
- Joakim Nivre, Uppsala University, Sweden;
- Eva Pettersson, Uppsala University, Sweden;
- Michael Piotrowski, University of Lausanne, Switzerland;
- Sophie Prévost, Laboratoire Lattice, France;
- Halim Sayoud, USTHB University;
- Olga Scrivner, Indiana University, USA;
- Neel Smith, College of the Holy Cross, USA;
- Sara Tonelli, Fondazione Bruno Kessler, Italy;
- Amir Zeldes, Georgetown University, USA;
- Daniel Zeman, Charles University, Czech Republic.