The research team from the Intelligent Agent System Lab (IASL), Institute of Information Science, Academia Sinica, Taiwan, led by Prof. Wen-Lian Hsu, and from Prof. Richard Tzong-Han Tsai's Intelligent Information Research Service Lab (IISR) at the Department of Computer Science & Engineering, Yuang Ze University, won the first place in the BioCreAtIvE II.5 Interactor Normalization Task Challenge, held in Madrid, Spain in 2009. There were 10 participating teams who submitted a total of 52 runs (33 offline and 19 online) to this Challenge. Other participants include teams from the Arizona State University, which won the first place last time in 2007, University of Tokyo, University of Colorado Denver and other renowned research teams from USA, Switzerland, Belgium, and Taiwan etc.

This Challenge is to evaluate the performance of state-of-the-art computer programs for the task of normalizing gene and gene product names in a corpus of full text articles. This task is an important step toward identifying protein-protein interaction in biomedical texts. Such computer programs can greatly reduce human curation effort for full-text articles. Comparing with other challenges, the new aspects include: 1) The use of an online server as basic evaluation infrastructure, allowing direct comparison of multiple participating systems, and 2) The use of full text articles other than abstract. The IASL-IISR team employed a high recall memory-based gene normalization (GN) approach, optimized for processing full text article, and a support vector machine based ranking method to achieve their top score.

IASL-IISR team members: Hong-Jie Dai, Po-Ting Lai, Chi-Hsin Huang, Yen-Ching Chang, Yue-Yang Bow, Hsin-Ta Wu, Prof. Richard Tzong-Han Tsai, and Prof. Wen-Lian Hsu