Room Assignment: Room 503
Participants will build systems to automatically detect the errors in Chinese sentences (error existence, types and position) made by Chinese-as-Second-Language learners. Four error types are annotated in the corpus for training, i.e. redundant word, missing word, word selection and word ordering. There were three previous events of the shared task (2014, 2015 and 2016).
Organized by: Gaoqi Rao, Baolin Zhang and Endong Xun
Contact: Gaoqi Rao, raogaoqi-fj@163.com
Given a word or phrase, participants are asked to provide a real-valued score from 1 to 9 for both valence and arousal dimensions, indicating the degree from most negative to most positive for valence, and from most calm to most excited for arousal.
Organized by: Liang-Chih Yu, Lung-Hao Lee, Jin Wang and Kam-Fai Wong
Contact: Liang-Chih Yu, lcyu@saturn.yzu.edu.tw
Participants will build systems to rank top-k reviews as a summary of opinions of product reviews in three different ways. The shared task will use a subset of Amazon SNAP product reviews dataset for experiments. It contains reviews of products in different categories, e.g., 22,000,000+ reviews of books and 7,000,000+ reviews of electronics products.
Organized by: Anil Kumar Singh, Julian McAuley, Avijit Thawani, Mayank Panchal, Anubhav Gupta and Rajesh Kumar Mundotiya
Contact: Anil Kumar Singh, aksingh.cse@iitbhu.ac.in
Understanding and being able to react to customer feedback is the most fundamental task in providing good customer service. The shared task provides corpora annotated using a five-plus-one-classes categorization (comment, request, bug, complaint, meaningless and undetermined) from a pilot study of ADAPT-Microsoft joint research project. Participants will train classifiers for the detection of meanings in customer feedback in English, French, Spanish and Japanese.
Organized by: Chao-Hong Liu, Declan Groves, Alberto Poncelas, Akira Hayakawa and Yasufumi Moriya
Contact: Chao-Hong Liu, chaohong.liu@adaptcentre.ie
Participants will build systems to choose the correct option for each multi-choice question. The dataset is comprised of two parts, English and Chinese. The total number of questions (along with their answers) is 14,447.
Organized by: Jun Zhao, Kang Liu, Shizhu He, Zhuoyu Wei and Shangmin Guo
Contact: mcqa.ijcnlp@outlook.com
Important Dates
Release of Training Data: May 15, 2017
Dryrun: Release of Development Set: July 12, 2017
Release of Test Set: August 14, 2017
Submission of Systems: August 21, 2017
System Description Paper Due: September 15, 2017
Camera-Ready Deadline: October 10, 2017
** All deadlines are calculated at 11:59pm UTC-7
Shared Task Co-Chairs
Chao-Hong Liu, ADAPT Centre, Dublin City University, chaohong.liu@adaptcentre.ie
Preslav Nakov, Qatar Computing Research Institute, pnakov@hbku.edu.qa
Nianwen Xue, Brandies University, xuen@brandeis.edu