SemEval-2018 Task 9: Hypernym Discovery

SemEval 2018 - Hypernym Discovery
The organizers of the SemEval-2018 – Task 9 are pleased to announce the shared task on Hypernym Discovery.
Hypernymy, i.e. the capability for generalization, lies at the core of human cognition. It is not surprising, therefore, that identifying hypernymic relations (i.e., is-a relations) has been pursued in NLP for over two decades. This task contributes in fact to the improvement of Question Answering, Textual Entailment and Semantic Search systems. In addition, hypernymic relations are the backbone of almost any ontology, semantic network and taxonomy, being therefore fundamental for downstream tasks, such as web retrieval, website navigation or records management.
In previous editions of SemEval, the identification of hypernymic relations from text corpora has been evaluated within the broader task of Taxonomy Evaluation. Most of other attempts have approached hypernymic relations through the task of Hypernym Detection, i.e. the binary task consisting of, given a pair of words, deciding whether a hypernymic relation holds between them or not. This experimental setting has already led to criticisms regarding its alleged oversimplification. Inspired by recent work, we propose therefore to reformulate the problem as Hypernym Discovery, i.e. given the search space of a domain’s vocabulary, and an input concept, discover its best (set of) hypernyms.
Colleagues are invited to participate by proposing a system for hypernym discovery by the evaluation deadline, on the 29th of January 2018 (test data will be released on the 8th of January). Training Data has just been released and it is accessible in the task’s website. Short papers describing the systems and their evaluation results must be submitted by the 26th of February 2018. All papers that meet the formal review criteria will be published in the workshop proceedings.
For further details and downloads see the task homepage:
For any communication, feel free to join our Google Group:
Important Dates:
25 Sep 2017: Training data release.
08 Jan 2018: Test data release. Evaluation start
29 Jan 2018: Evaluation end
05 Feb 2018: Results Posted
26 Feb 2018: System description paper submissions due by 23:59 GMT -12:00
19 Mar 2018: Paper reviews due
02 Apr 2018: Author notifications
16 Apr 2018: Camera ready submissions due
Shared Task Organizers:
Jose Camacho-Collados – Sapienza University of Rome
Claudio Delli Bovi – Sapienza University of Rome
Tommaso Pasini – Sapienza University of Rome
Roberto Navigli – Sapienza University of Rome
Vered Shwartz – Bar-Ilan University
Luis Espinosa-Anke – Universitat Pompeu Fabra
Sergio Oramas – Universitat Pompeu Fabra
Horacio Saggion – Universitat Pompeu Fabra

Related posts

Leave a Comment