2008.1.6-11 インド・ハイデラバード
The Third International Joint Conference on Natural Language Processing (IJCNLP08) 自然言語処理学講座: 博士後期課程3年 |
|
【発表内容】 Aiming at acquiring semantic relations between events from a large corpus, this paper proposes several extensions to a state-of-the-art method originally designed for entity relation extraction, reporting on the present results of our experiments on a Japanese Web corpus. The results show that (a) there are indeed specific cooccurrence patterns useful for event relation acquisition, (b) the use of cooccurrence samples involving verbal nouns has positive impacts on both recall and precision, and (c) over five thousand relation instances are acquired from a 500M-sentence Web corpus with a precision of about 66% for action-effect relations. 【会議の内容】 著名な研究者による講演を楽しみにしていたが、それほどのインパクトはなかった。 ポスターセッションの"A re-examination of dependency path kernels for relation extraction"は私の研究に応用できそうな内容だった。手法はよいが 結果は悪かった。そのため、私の問題に応用するのは難しそうだった。また、 もっとシンプルな手法でも良い成果を挙げられるとの報告があるので、彼の問 題にこの手法が有効なのかは疑問であった。いくつか問題があるにせよ、面白 い研究を知ることができたので満足しました。 【 他の参加者との研究技術交流等】 昼食の間は他の研究者と沢山意見交換をしました(昼食もおいしかったです)。 |
|
2008.1.6-12 インド・ハイデラバード
The Third International Joint Conference on Natural Language Processing (IJCNLP08) The Sixth SIGHAN Workshop on Chinese Language Processing 自然言語処理学講座: 博士前期課程2年 |
|
【発表内容】 In our paper, we first categorize Chinese synthetic words into several types according to their inside semantic and syntactic structure, and then propose a method to represent these inside information of word by applying a tree-based structure. Then we try to automatically identify the inner morphological structure of 3-character synthetic words by using a large corpus and try to add syntactic tags to their internal structure. We believe that this tree-based word internal information could be useful in specifying a Chinese synthetic word segmentation standard. 【会議の内容】 It was a high level conference which had a lot of interesting papers and demos. Here are some of the papers that I felt interested on them. “Automatic rule acquisition for Chinese intra-chunk relations, Qiang Zhou, Tsinghua University, Beijing” This paper proposed a approach to automatically extract rules from Chinese word chunks which are the same target with our own research. However, it chosen a different way to analysis the internal structure of Chinese words which is quite useful to our research. “An Empirical Comparison of Goodness Measures for Unsupervised Chinese Word Segmentation with a Unified Framework, Hai Zhao and Chunyu Kit” This paper compared goodness measures for unsupervised Chinese word segmentation, The result shows the system that use character based method has the best performance until now. And the authors proposed their own method using CRF rather than any other machine learning method and gain a better result at last. 【研究技術交流等】 I had a discussion with Prof. Qiang Zhou, who comes from Tsinghua University, on what is the best way to categories the internal structure of Chinese synthetic words. And I gained his permission on using their corpus for our own research on Chinese synthetic words. |
|
2008.1.7-13 インド・ハイデラバード
The Third International Joint Conference on Natural Language Processing (IJCNLP08) 自然言語処理学講座: 博士後期課程3年 |
|
|