The International FrameNet Workshop 2020: Towards a Global, Multilingual FrameNet, collocated with LREC, in Marseille, France, will bring together researchers in Frame Semantics and Construction Grammar, two areas which have traditionally been interrelated, but which have been developing somewhat independently in recent years. It is also addressed at language technology researchers working with language resources based on Frame Semantics or Construction Grammar. IFNW 2020 follows from three previous editions, one in 2018, held in Miyazaki, Japan, one in 2016, held in Juiz de Fora, Brazil, and one in 2013, held in Berkeley, USA.

Important dates

• Paper Submission: Feb. 21st, 2020 (GMT-12)

• Author notification: March 13th, 2020

• Camera-ready papers due: April 8th, 2020 (GMT-12)

• Workshop: June 2nd, 9th, 16th, 23rd and 30th.

Program

The International FrameNet Workshop 2020 will be hosted online (instructions will be provided to registered participants in due time), on every Tuesday of June 2020. There will be a total of five sessions, scheduled as follows:

June 2nd, 10am - 11am (UTC)

June 9th, 4pm - 5pm (UTC)

June 16th, 12pm - 2pm (UTC)

Posters

• Combining Conceptual and Referential Annotation to Study Variation in Framing (Postma, Remijnse, Ilievski, Fokkens, Titarsolej & Vossen)

• FrameNet Annotations Alignment using Attention-based Machine Translation (Marzinotto)

• Greek within the Global FrameNet Initiative: Challenges and Conclusions so far (Giouli, Pilitsidou & Christopoulos)

• Using Verb Frames for Text Difficulty Assessment (Lee, Liu & Cai)

• Deriving a PropBank Corpus from Parallel FrameNet and UD Corpora (Gruzitis, Dargis, Rituma, Nešpore-Berzkalne & Saulite)

• Building the Emirati Arabic FrameNet (Gargett & Leung)

June 23rd, 9pm - 10pm (UTC)

June 30th, 2pm - 3pm (UTC)

Discussion Session

Call for papers

Starting from the original FrameNet project (for English) at the International Computer Science Institute (ICSI), projects for lexical resources based on Frame Semantics have sprouted in more than a dozen countries, with major efforts including Spanish, Japanese, Swedish, Brazilian Portuguese, Chinese, German, French, Korean, Dutch, Latvian, Finnish, and Hebrew, and a variety of other languages. Continuing the series of International FrameNet Workshops in 2013, 2016, 2017, and 2018, this one-day workshop will promote the exchange of ideas, data,  techniques, and software among these projects and their many users. We welcome papers discussing the differences and similarities between work on Frame Semantics in different languages from both theoretical and practical perspectives.  

The theoretical papers might include questions such as:

• What determines which frames are similar (or essentially identical) across languages? How can we characterize the differences between frames across languages?

• What can we learn from parallel annotation in general and the current TED talk parallel annotation in particular?  How does that relate to translation theory and practice?

• What can we learn from or contribute to other meaning representations such as PropBank, AMR, UCCA predicate logic, etc.?  How can we integrate frames for general vocabulary with those for specialized domains? Are there semantic domains where Frame Semantics does not seem applicable?

Practical papers might discuss questions such as:

• How can we improve collaboration between Frame Semantic projects for different languages? Are there methods to design databases, create software, do annotation, etc. that will facilitate reuse, especially by new projects? What role could/should/do machine learning and machine translation play in developing FrameNets?

• What policies regarding public release of data are in place for each project?  Should we aim at a common policy?  Should there be limitations on the use of FrameNet data? If so, what?  What role does/could commercial support play in your work? 

• What relation is there now or should there be between projects which are primarily based on Frame Semantics and those primarily based on Construction Grammar? Between Frame Semantic projects and those based on other meaning representations or aggregations of representations, such as Uby, BabelNet, and Framester?

Part of the meeting will be devoted to presentation of oral papers and posters. However, given the very different situations of the various projects, we will also allow ample time for small group and informal discussions, which are often the best way to promote mutual understanding and cooperation, and to resolve practical questions.

All submissions must follow the LREC formatting guidelines and be submitted via the START page. When submitting a paper, authors will be asked to provide essential information about resources (in a broad sense, i.e. also technologies, standards, evaluation kits, etc.) that have been used for the work described in the paper or are a new result of your research. Moreover, ELRA encourages all LREC authors to share the described LRs (data, tools, services, etc.) to enable their reuse and replicability of experiments (including evaluations).