| 札場 寛之 | 1551094: M, 2回目発表 | 知能コミュニケーション | 中村 哲,松本 裕治,Graham Neubig,吉野 幸一郎 |
|
title: Generating Source Code from Natural Language using Statistical Machine Translation
abstract: We propose a method to generate a program source code from natural language using statistical machine translation. In order to generate a syntactically correct source code, we generate abstract syntax tree(AST). We take care of implicit restriction which AST has using some rules. language of the presentation: Japanese | |||
| 前田 若菜 | 1551100: M, 2回目発表 | 知能コミュニケーション | 中村 哲,松本 裕治,鈴木 優,吉野 幸一郎,Graham Neubig |
|
title: Anonymization technique for Unstructured text data considering inference from context
abstract: It needs anonymization to open unstructurized text data wihich includes personal information. Conventional redaction methods using a reference list have a problem that the number of remaining characters in raw data decreases. For solving it, we propose a k-anonymity method for the substring of target character string to be anonymized based on pattern-matching. language of the presentation: Japanese | |||
| 森下 睦 | 1551107: M, 2回目発表 | 知能コミュニケーション | 中村 哲,松本 裕治,Graham Neubig,吉野 幸一郎 |
|
title: Parser Self-Training for Syntax-Based Machine Translation
abstract: In syntax-based machine translation, it is known that the accuracy of parsing greatly affects the translation accuracy. Self-training, which uses parser output as training data, is one method to improve the parser accuracy. However, because parsing errors cause noisy data to be mixed with the training data, automatically generated parse trees do not always contribute to improving accuracy. We propose a method for selecting self-training data by performing syntax-based machine translation using a variety of parse trees, using automatic evaluation metrics to select which translation is better, and using that translation's parse tree for parser self-training. Our self-trained parsers improve a state-of-the-art translation system in two language pairs. language of the presentation: Japanese | |||