The international conference series Finite-State Methods and Natural Language Processing (FSMNLP) is the premier forum of the ACL Special Interest Group on Finite-State Methods (SIGFSM). It serves researchers and practitioners working on natural language processing (NLP) applications or language resources, theoretical and implementational aspects, or their combinations, that make use of finite-state methods.
FSMNLP 2017 takes place in Umeå, Sweden, and is co-located with the 13th International Workshop on Tree Adjoining Grammars and Related Formalisms (TAG+13), fostering the scientific exchange between the two research communities. The meeting will include a number of joint tutorial presentations, designed to introduce attendees to a range of topics relevant to both communities.
FSMNLP 2017 invites papers on natural language processing (NLP) with obvious relevance or an explicit relation to finite-state methods as well as, conversely, papers on finite-state methods that are motivated by or clearly relevant to NLP. Topics include but are not limited to:
- NLP applications and linguistic aspects of finite-state methods
- Finite-state models of natural language and linguistic theories
- Practices for building morphological models using finite-state technology
- Machine learning of finite-state models of natural language
- Finite-state manipulation software and tools with relevance to NLP
- Practical implementations of linguistic descriptions with finite-state technology, including grammars, machine learning tools, and language-specific challenges to finite-state NLP
- Formal results with relevance to finite state machines and description languages
- Applications of finite-state-based NLP in fields such as comparative linguistics, field linguistics, applied linguistics and language teaching.
FSMNLP 2017 Special Theme
FSMNLP 2017 features the special theme
Finite-State Methods for Semantic Structures
covering, e.g., the creation, analysis, transformation of semantic structures of various descriptions, learning and training methods for finite-state devices processing semantic structures, etc.
The special theme is in no way intended to restrict the scope of papers. Submissions relating to any dimension of finite-state NLP are encouraged, whether or not they belong to the special theme. In particular, this includes papers which apply finite-state methods to new domains beyond the traditional focus of FSMNLP.
|June 11, 2017||Extended paper submission deadline|
|July 16, 2017||Notification of acceptance|
|August 6, 2017||Camera-ready version due|
|September 4-6, 2017||Conference|
Deadlines are midnight Swedish time (UTC+2).
Papers must present original, unpublished research and implementation results. Simultaneous submissions to FSMNLP and to another conference or workshop are allowed. According to the ACL double submission policy (https://www.aclweb.org/archive/policies/current/double-submission-policy.html), doubly-submitted papers should indicate this fact on the title page. If accepted at both FSMNLP and the other venue, such a paper must be withdrawn from one of them. Withdrawal from FSMNLP shall be as soon as possible but no later than one week after the notification of acceptance at FSMNLP.
FSMNLP accepts two kinds of submissions:
long papers (8 pages excluding references) reporting completed, significant research,
short papers (4 pages excluding references) reporting ongoing work and partial results, implementations, grammars, practical tools, interactive software demos, etc.
Short papers are expected to be presented as system demos, posters and/or short presentations, while long papers are presented in full talks.
All submissions are electronic and in PDF format via a web-based submission server. Authors are strongly encouraged to preparte their papers in LaTeX using the ACL 2016 style files to produce the PDF document. The zip archive of the style files can be downloaded here:
Information about the author(s) and other identifying information such as obvious self-references (e.g., “We showed in  …”) and financial or personal acknowledgements should be omitted in the submitted papers whenever feasible.
Papers must be submitted electronically in PDF at
A paper may contain a clearly marked appendix and data files to support its claims. This material will not be published. While reviewers are urged to consult this extra material for better comprehension, it is at their discretion whether they do so. Such extra material should also be anonymized to the extent feasible.
All accepted papers will be published and archived in the FSMNLP 2017 proceedings in the ACL Anthology. We plan to provide the opportunity to submit revised and substantially extended versions of accepted papers to a special issue of an appropriate high-quality journal.