| Presentation | 2007/11/21 Open-Vocabulary Spoken Utterance Retrieval Using Confusion Networks and Its Evaluation Takaaki HORI, I. Lee HETHERINGTON, Timothy J. HAZEN, James R. GLASS, |
|---|---|
| PDF Download Page | PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | We recently proposed an open-vocabulary spoken utterance retrieval method using confusion networks. In this paper, we particularly focus on its technical details and experimental results with a large-scale corpus of approximately 120 hours. Spoken utterance retrieval is a task to find utterances including a key word or key phrase as a query from a large spoken archive. The proposed method generates word or phone confusion networks from the archive, and translates them into the Inverted Index. By using this index structure, we can quickly retrieve a set of utterances including the query. To deal with out-of-vocabulary (OOV) words in queries and the archive, we apply phone confusion networks and combine them with word confusion networks. With this approach, we enables robust keyword matching for queries including both in-vocabulary and out-of-vocabulary words. In the retrieval experiments with speech recordings in MIT lecture corpus, our method using word-phone-combined confusion networks outperformed conventional indexing and retrieval methods. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Open Vocabulary / Utterance Retrieval / Confusion Network |
| Paper # | SP2007-93 |
| Date of Issue |
| Conference Information | |
| Committee | SP |
|---|---|
| Conference Date | 2007/11/21(1days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | |
| Vice Chair | |
| Secretary | |
| Assistant | |
| Paper Information | |
| Registration To | Speech (SP) |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Open-Vocabulary Spoken Utterance Retrieval Using Confusion Networks and Its Evaluation |
| Sub Title (in English) | |
| Keyword(1) | Open Vocabulary |
| Keyword(2) | Utterance Retrieval |
| Keyword(3) | Confusion Network |
| 1st Author's Name | Takaaki HORI |
| 1st Author's Affiliation | NTT Communication Science Laboratories, NTT Corporation() |
| 2nd Author's Name | I. Lee HETHERINGTON |
| 2nd Author's Affiliation | MIT Computer Science and Artificial Intelligence Laboratory |
| 3rd Author's Name | Timothy J. HAZEN |
| 3rd Author's Affiliation | MIT Computer Science and Artificial Intelligence Laboratory |
| 4th Author's Name | James R. GLASS |
| 4th Author's Affiliation | MIT Computer Science and Artificial Intelligence Laboratory |
| Date | 2007/11/21 |
| Paper # | SP2007-93 |
| Volume (vol) | vol.107 |
| Number (no) | 356 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |