| Project/Area Number |
11680386
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | Nagoya University |
|---|
Principal Investigator |
TAKEDA Kazuya Graduate School of Engineering, Nagoya University, Associate Professor, 工学研究科, 助教授 (20273295)
|
|---|
| Project Period (FY) |
1999 – 2001
|
| Project Status |
Completed (Fiscal Year 2001)
|
| Budget Amount *help |
¥3,100,000 (Direct Cost: ¥3,100,000)
Fiscal Year 2001: ¥1,200,000 (Direct Cost: ¥1,200,000)
Fiscal Year 2000: ¥1,000,000 (Direct Cost: ¥1,000,000)
Fiscal Year 1999: ¥900,000 (Direct Cost: ¥900,000)
|
|---|
| Keywords | Large corpus / Morphological analysis / Continuous speech recognition / Language model / Acoustic model / Entropy / 音声認識 / 音韻モデル / 特異値分解 / Nグラム |
|---|
| Research Abstract |
(1)A new formula for calculating the probabilities of the given word sequence arises. The new probabilistic model of language takes the length of the word sequence into account.
(2) It is found that the insertion penalty in combining the language and the acoustic probabilities compensates the difference of the model and true entropy of the language.
(3) A linear space representation of trigram probability is proposed for interpolating the probabilities and for clustering words.
|
