2008Q3 Reports: SIGHAN

2007-08 SIGHAN Report

1. SIGHAN-6 Workshop

The 6th SIGHAN Workshop was held in conjunction with the 3rd International Joint Conference on Natural Language Processing (IJCNLP 2008) in Hyderabad, India, 11-12 January 2008. This is the second time for which a SIGHAN workshop co-locates with IJCNLP, the flagship conference of the Asian Federation of Natural Language Processing.

The workshop was co-chaired by Olivia Kwong from the City University of Hong Kong and Haizhou Li from the Institute for Infocomm Research in Singapore. A total of 9 submissions were accepted for presentations, contributing to yet another stimulating and inspiring scientific program which covers Chinese word segmentation, parsing, coreference resolution, temporal relation identification, and machine transliteration, amongst others.

SIGHAN-6 also uniquely features the Fourth International Chinese Language Processing Bakeoff, which was jointly held with the First CIPS Chinese Language Processing Evaluation over the summer of 2007. To promote this event, a special panel on "Chinese Word Segmentation Evaluation and Related Issues", jointly held by SIGHAN and the Chinese Information Processing Society of China, and endorsed by the Asian Federation of Natural Language Processing, was held in Dalian on 6 August 2007. Invited panelists include Chang-Ning Huang from Microsoft Research Asia, Benjamin Tsou from the City University of Hong Kong, Gina-Anne Levow from the University of Chicago, Guangjin Jin from the Institute of Applied Linguistics, MOE, China, and Qun Liu from the Chinese Academy of Sciences. The panel discussion was moderated by Olivia Kwong from the City University of Hong Kong.

The bakeoff was co-organized by SIGHAN, ChineseLDC, and the Verifying Center on Chinese Language and Character Standards of the State Language Commission of PRC, and coordinated by Dr. Guangjin Jin of the Institute of Applied Linguistics, MOE, China. In addition to the classic Chinese word segmentation and named entity recognition tasks, there is a new track on Chinese POS-tagging. Benchmarking corpora were generously provided by Academia Sinica; City University of Hong Kong; Institute of Applied Linguistics, MOE, China; Microsoft Research Asia; Peking University; Shanxi University; and University of Colorado. There were 28 teams participating in the bakeoff, and 23 system reports were collected in the proceedings.

Six of the regular papers were selected and invited for further revision and publication in a special issue of the International Journal of Computer Processing of Oriental Languages (IJCPOL), Vol.21, No.1, to appear in April/May 2008.

2. SIGHAN Organisation

The SIGHAN constitution has been slightly revised to allow a member-at-large to be elected from Asia. In this regard, Qing Ma from Ryukoku University was elected as the member-at-large. It is also noted that Dan Jurafsky will take over from Benjamin Tsou as the Chair of SIGHAN from the next term of office.