智能技术与自然语言处理实验室将于本周五（5月27号）晚6：00在新技术楼618会议室举行学术活动，会上由博士生刘桃、赵玉茗为大家作报告，请全体师生准时到会参加！ 报告内容如下：
1. Liu Tao
Title: A robust statistical method for unknown word detection
Abstract: Word segmentation for many Asian languages is a prerequisite step for many NLP tasks. The existence of out of vocabulary (OOV) makes word segmentation more difficult. Especially for large scale web texts, OOV problem is more distinct. An efficient and robust method to detect unknown word is needed greatly. This paper proposes a fast and robust unknown word detection method. It is based on fast n-gram computing method, local maxima algorithm and other heuristic features. Experiment on sighan bakeoff corpora shows the effectiveness of this method.
2. Zhao YuMing
Title: Structure of QA System And Sentence Similarity Calculation Including Question Analyzing Information
Abstract: Structure of our QA System will be introduced in this presentation. And the realization of this system will also be reported. Besides that, an approach of sentence similarity calculation including question analyzing information will be presented.