tailieunhanh - Báo cáo khoa học: "Using Generation for Grammar Analysis and Error Detection"

We demonstrate that the bidirectionality of deep grammars, allowing them to generate as well as parse sentences, can be used to automatically and effectively identify errors in the grammars. The system is tested on two implemented HPSG grammars: Jacy for Japanese, and the ERG for English. Using this system, we were able to increase generation coverage in Jacy by 18% (45% to 63%) with only four weeks of grammar development. | Using Generation for Grammar Analysis and Error Detection Michael Wayne Goodman University of Washington Dept. of Linguistics Box 354340 Seattle Wa 98195 USA goodmami@ Francis Bond NICT Language Infrastructure Group 3-5 Hikaridai Seika-cho Soraku-gun Kyoto 619-0289 Japan bond@ Abstract We demonstrate that the bidirectionality of deep grammars allowing them to generate as well as parse sentences can be used to automatically and effectively identify errors in the grammars. The system is tested on two implemented HPSG grammars Jacy for Japanese and the ERG for English. Using this system we were able to increase generation coverage in Jacy by 18 45 to 63 with only four weeks of grammar development. 1 Introduction Linguistically motivated analysis of text provides much useful information for subsequent processing. However this is generally at the cost of reduced coverage due both to the difficulty of providing analyses for all phenomena and the complexity of implementing these analyses. In this paper we present a method of identifying problems in a deep grammar by exploiting the fact that it can be used for both parsing interpreting text into semantics and generation realizing semantics as text . Since both parsing and generation use the same grammar their performance is closely related in general improving the performance or cover of one direction will also improve the other. Flickinger 2008 The central idea is that we test the grammar on a full round trip parsing text to its semantic representation and then generating from it. In general any sentence where we cannot reproduce the original or where the generated sentence significantly differs from the original identifies a flaw in the grammar and with enough examples we can pinpoint the grammar rules causing these problems. We call our system Egad which stands for Erroneous Generation Analysis and Detection. This research was carried out while visiting NICT. 2 Background This work was inspired .

• Báo cáo khoa học
• Using Generation for Grammar Analysis
• Error Detection
• Michael Wayne
• báo cáo khoa học
• báo cáo ngôn ngữ
• xử lý ngôn ngữ tự nhiên
• Bài giảng Xử lý ngôn ngữ tự nhiên
• Ngôn ngữ tự nhiên
• Kỹ thuật lập trình
• Xử lý nhập nhằng ngữ nghĩa
• Xử lý ngữ nghĩa
• Xử lý ngôn ngữ
• Natural Language Processing
• Ngôn ngữ lập trình
• Nhập nhằng cấu trúc
• Tri thức về ngôn ngữ
• Mô hình ngôn ngữ
• Mô hình n gram
• Dịch máy
• Phương pháp dịch máy
• Hiểu ngôn ngữ
• Phân tích ngữ nghĩa
• Phân loại tin tự động
• Biểu diễn vị từ
• Thuộc tính về sự kiện
• Thuộc tính về sự kiện
• Mô hình ngôn ngữ Google Book N grams
• Mô hình ngôn ngữ KenLM
• Phương pháp làm mịn
• Tách từ tiếng Việt
• Gán nhãn từ loại
• Phân tích cú pháp
• Nghĩa từ vựng
• Phân giải nhập nhằng từ
• Phân loại văn bản
• Lexical chain
• Phản hồi thông tin
• Bài toán PTCP
• Cấu trúc ngữ pháp
• Phân tích cú pháp xác suất
• CKY kết hợp xác suất
• Văn phạm phi ngữ cảnh xác suất
• Hình thái học