Method of estimating probabilities of occurrence of speech vocabulary elements

• US 6,314,400 B1
• Filed: 09/14/1999
• Issued: 11/06/2001
• Est. Priority Date: 09/16/1998
• Status: Expired due to Fees

• Current Assignee: US Philips Corporation (Koninklijke Philips N.V.)
• Original Assignee: US Philips Corporation (Koninklijke Philips N.V.)
• Inventors: Klakow, Dietrich
• Primary Examiner(s): {haeck over (S)}mits, Ta̅livaldis Ivars
• Assistant Examiner(s): ABEBE, DANIEL DEMELASH
• Application Number:
  US09/396,086
• Time in Patent Office: 784 Days
• Field of Search: 704/257, 704/231, 704/232, 704/240, 704/241, 704/243, 704/251, 704/256, 704/250, 704/255
• US Class Current: 704/257
• CPC Class Codes: G10L 15/197 Probabilistic grammars, e.g. word n-grams

Abstract

The invention relates to a method of estimating probabilities of occurrence of speech vocabulary elements in a speech recognition system. By modification of the linguistic speech modeling, further alternatives for reducing the error rate and perplexity of a speech recognition system are proposed. The method according to the invention is characterized in that, in the estimation of a probability of occurrence of a speech vocabulary element, several M-gram probabilities of this element are raised to a higher power by means of an M-gram-specific optimized parameter value, and the powers thus obtained are multiplied by each other, in which the estimation of the probability of occurrence of a speech vocabulary element does not include the case where an M-gram probability with M>1 estimated by means of a first training vocabulary corpus for the speech vocabulary element is multiplied by a quotient raised to a higher power by means of an optimized parameter value, which optimized parameter value is determined by means of the GIS algorithm, and a unigram probability of the element estimated by means of a second training vocabulary corpus serves as a dividend of the quotient, and a unigram probability of the element estimated by means of the first training vocabulary corpus serves as a divisor of the quotient.

Claims

1. A method of estimating probabilities of occurrence of speech vocabulary elements in a speech recognition system, wherein, in the estimation of a probability of occurrence of a speech vocabulary element, several M-gram probabilities of this element are raised to a higher power by means of an M-gram-specific optimized parameter value, and the powers thus obtained are multiplied by each other, in which the estimation of the probability of occurrence of a speech vocabulary element does not include the case where an M-gram probability with M>

1 estimated by means of a first training vocabulary corpus for the speech vocabulary element is multiplied by a quotient raised to the power of an optimized parameter value, which optimized parameter value is determined by means of the GIS algorithm, and a unigram probability of the element estimated by means of a second training vocabulary corpus serves as a dividend of the quotient, and a unigram probability of the element estimated by means of the first training vocabulary corpus serves as a divisor of the quotient.

2. A method as claimed in claim 1, wherein a first training vocabulary corpus is used for estimating a first part of the M-gram probabilities, and a first part of a second training vocabulary corpus is used for estimating a second part of the M-gram probabilities, and a second part of the second training vocabulary corpus is used for determining the optimized parameter values assigned to the M-gram probabilities.

3. A method as claimed in claim 2, wherein, for determining the optimized parameter values, the optimizing function $F$

({λ

i})=∑

hw

f

(hw)

log

 



(1Zλ



(h)

∏

i

pi(w



h)λ

i)

is minimized, wherein

λ

_i represents the parameter values to be optimized,

hw represents M-grams for a vocabulary element w with a history h of previous vocabulary elements,

f(hw) represents the quotient with the number of counted M-grams occurring in a training phase of the second part of the second vocabulary as a dividend, and the number of vocabulary elements of the second vocabulary as a divisor,

1/Z_λ

(h) represents a scaling factor, and

pi represents the estimated probability of occurrence of the vocabulary element w, given the history h.

4. A method as claimed in claim 1, wherein only M-gram probabilities with M<

3 are used.

5. A method as claimed in claim 4, wherein gap bigram probabilities are used.

6. A speech recognition system using a speech vocabulary with vocabulary elements to which probabilities of occurrence are assigned, which are estimated by means of a method as claimed in claim 1.