CA2384518A1 - Determining a compact model to transcribe the arabic language acoustically in a well defined basic phonetic study - Google Patents
Determining a compact model to transcribe the arabic language acoustically in a well defined basic phonetic study Download PDFInfo
- Publication number
- CA2384518A1 CA2384518A1 CA002384518A CA2384518A CA2384518A1 CA 2384518 A1 CA2384518 A1 CA 2384518A1 CA 002384518 A CA002384518 A CA 002384518A CA 2384518 A CA2384518 A CA 2384518A CA 2384518 A1 CA2384518 A1 CA 2384518A1
- Authority
- CA
- Canada
- Prior art keywords
- language
- transcribe
- well
- phonetics
- compact model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Abandoned
Links
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
- G10L15/18—Speech classification or search using natural language modelling
- G10L15/183—Speech classification or search using natural language modelling using context dependencies, e.g. language models
- G10L15/187—Phonemic context, e.g. pronunciation rules, phonotactical constraints or phoneme n-grams
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/02—Feature extraction for speech recognition; Selection of recognition unit
- G10L2015/025—Phonemes, fenemes or fenones being the recognition units
Abstract
In the development of an automatic speech recognition (ASR) system, an extensive study of the basic phonetic alphabet is performed to collect information regarding phonology and phonetics of the language or dialect in question (modem standard Arabic or MSA in this case). In addition, terminological and transcriptional problems are identified with respect to the language or dialect in question.
Next, based on feature description (rather than symbol shapes), the symbols in the literature are mapped to a single or more recent phonetic alphabet. Lastly, from a maximal set containing all the phonemes, allophones, and transliteration symbols, a reduced set is created with a compact set of phonetic alphabets. Memory consumption is greatly reduced in a computer system by using this compact set of phonetic alphabets
Next, based on feature description (rather than symbol shapes), the symbols in the literature are mapped to a single or more recent phonetic alphabet. Lastly, from a maximal set containing all the phonemes, allophones, and transliteration symbols, a reduced set is created with a compact set of phonetic alphabets. Memory consumption is greatly reduced in a computer system by using this compact set of phonetic alphabets
Description
TITLE:
~'e,-wrpa6,r~,41 ~ ' t.
~ .~ ~ ~G' ~ ~' to.
Determining a Compact Model to Transcribe the Arabic Language Acoustically in a Well Defined Basic Phonetic Study Field of Invention The present invention relates generally to the field of controlling a computer dictation application using multi-gender human voice instead of a keyboard.
More specifically, the present invention is related to determining a compact model to transcribe the Arabic language acoustically in a well-defined basic phonefic study.
Background of the Invention Phonetics, as defined by the Merriam-Webster~ dictionary (Collegiate 10th ed.), is a system of speech sounds of a language or group of languages, and further comprises the study and systematic classification of the sounds made in spoken utterance. Hence, the phonetic system represents the practical application of this science to language study. An important part of phonetics is phonemes.
Phonemes, as defined by Merriam-Webster~ dictionary (Collegiate 10th ed.), are abstract units of the phonetics system (associated with a particular language) that correspond to a group of speech sounds. For example, velar ~k~ of cry/ and the palatal ~k~ of keel are distinct sounds in the English language and are part of a set similar speech sounds. Another term related to Phonemes is allophones.
Allophone, as defined by Merriam-Webster° dictionary (Collegiate 10~' ed.), is one of two or more variants of the same phoneme. For example, the aspirated (p~ of pin and the unaspirated ~p~ of spin are allophones of the phoneme ~p~.
Express Mail L,ebd wt. ~, ~ caa~R209US
~~2/ A05;p Orthography is another system associated with the sounds of a given language.
Orthography, as defined by Merriam-Webster~ dictionary (Collegiate 10~' ed.), is the representation of the sounds of a language by letters and diacritics. A
diacritic is further defined as a mark near or through an orthographic or phonetic character or combination of characters indicating a phonetic value different from that given the unmarked or otherwise marked element. An example of a diacritic is the acute accents of resume, which are added to the letter a to indicate a special phonetic value.
Additionally, some foreign languages often use diacritics to double the force of the phoneme, and they further use geminated graphemes. Graphemes are the set of units of a writing system (as letters and letter combinations) that represent a phoneme.
Geminated graphemes are a sequence of identical speech sounds (as in meanness or Italian none).
One of the advances in recent years is the impact of computers in the field of phonetics. One of the major challenges associated with human speech and computers is automatic speech recognition or ASR. ASR is defined as the ability of a computer-based system to recognize and decipher human voice. ASR systems are usually programmed to recognize a simple set of words that are common to a group of users, or sometimes ASR systems are programmed to recognize a complex set of words associated with a specific user.
One common problem associated with phonetic representation of foreign language (such as Arabic) is the abundance of phonetics associated with such languages. Arabic language displays a difference in orthography and phonetics
~'e,-wrpa6,r~,41 ~ ' t.
~ .~ ~ ~G' ~ ~' to.
Determining a Compact Model to Transcribe the Arabic Language Acoustically in a Well Defined Basic Phonetic Study Field of Invention The present invention relates generally to the field of controlling a computer dictation application using multi-gender human voice instead of a keyboard.
More specifically, the present invention is related to determining a compact model to transcribe the Arabic language acoustically in a well-defined basic phonefic study.
Background of the Invention Phonetics, as defined by the Merriam-Webster~ dictionary (Collegiate 10th ed.), is a system of speech sounds of a language or group of languages, and further comprises the study and systematic classification of the sounds made in spoken utterance. Hence, the phonetic system represents the practical application of this science to language study. An important part of phonetics is phonemes.
Phonemes, as defined by Merriam-Webster~ dictionary (Collegiate 10th ed.), are abstract units of the phonetics system (associated with a particular language) that correspond to a group of speech sounds. For example, velar ~k~ of cry/ and the palatal ~k~ of keel are distinct sounds in the English language and are part of a set similar speech sounds. Another term related to Phonemes is allophones.
Allophone, as defined by Merriam-Webster° dictionary (Collegiate 10~' ed.), is one of two or more variants of the same phoneme. For example, the aspirated (p~ of pin and the unaspirated ~p~ of spin are allophones of the phoneme ~p~.
Express Mail L,ebd wt. ~, ~ caa~R209US
~~2/ A05;p Orthography is another system associated with the sounds of a given language.
Orthography, as defined by Merriam-Webster~ dictionary (Collegiate 10~' ed.), is the representation of the sounds of a language by letters and diacritics. A
diacritic is further defined as a mark near or through an orthographic or phonetic character or combination of characters indicating a phonetic value different from that given the unmarked or otherwise marked element. An example of a diacritic is the acute accents of resume, which are added to the letter a to indicate a special phonetic value.
Additionally, some foreign languages often use diacritics to double the force of the phoneme, and they further use geminated graphemes. Graphemes are the set of units of a writing system (as letters and letter combinations) that represent a phoneme.
Geminated graphemes are a sequence of identical speech sounds (as in meanness or Italian none).
One of the advances in recent years is the impact of computers in the field of phonetics. One of the major challenges associated with human speech and computers is automatic speech recognition or ASR. ASR is defined as the ability of a computer-based system to recognize and decipher human voice. ASR systems are usually programmed to recognize a simple set of words that are common to a group of users, or sometimes ASR systems are programmed to recognize a complex set of words associated with a specific user.
One common problem associated with phonetic representation of foreign language (such as Arabic) is the abundance of phonetics associated with such languages. Arabic language displays a difference in orthography and phonetics
2 roz~ ~~~~~
associated with the language. This is best illustrated by the example of geminated graphemes. The feature of gemination is inherent in most Arabic phonetic alphabets, which is defined as doubling the force of the phoneme that is marked by a superscript sign. During editing, people do not write this sign unless it is crucially needed to decipher a certain meaning from another. That's why the grapheme is written only once. Another example is that the language exhibits different variations between the kinds of vowel distribution; either they are short or long vowels. Tables 2 and 3, as detailed hereafter, represent the different features of vowels and gemination consecutively. Thus, software representing such a system comprising a myriad of phonetics inevitably requires a significant allocation of memory on a computer-based device for storage of such plurality of phonetics.
A variety of software applications are available today that utilize the phonetics system to recognize the speech of human users. But, none of the prior art software utilizes an automatic speech recognition system that uses an orthographic system comprising a compact set of phonetics. Whatever the precise merits, features and advantages of the above cited references, none of them achieves or fulfills the purposes of the present invention.
~~1MMARY OF THE INVENTION
The present invention provides for a method and a system for developing a compact model to transcribe the Arabic language acoustically based on a well-defined basic phonetic study. The compact model is accomplished in the present invention by reducing the set of phonemes. Thus, the creation of a minimized set of phonemes
associated with the language. This is best illustrated by the example of geminated graphemes. The feature of gemination is inherent in most Arabic phonetic alphabets, which is defined as doubling the force of the phoneme that is marked by a superscript sign. During editing, people do not write this sign unless it is crucially needed to decipher a certain meaning from another. That's why the grapheme is written only once. Another example is that the language exhibits different variations between the kinds of vowel distribution; either they are short or long vowels. Tables 2 and 3, as detailed hereafter, represent the different features of vowels and gemination consecutively. Thus, software representing such a system comprising a myriad of phonetics inevitably requires a significant allocation of memory on a computer-based device for storage of such plurality of phonetics.
A variety of software applications are available today that utilize the phonetics system to recognize the speech of human users. But, none of the prior art software utilizes an automatic speech recognition system that uses an orthographic system comprising a compact set of phonetics. Whatever the precise merits, features and advantages of the above cited references, none of them achieves or fulfills the purposes of the present invention.
~~1MMARY OF THE INVENTION
The present invention provides for a method and a system for developing a compact model to transcribe the Arabic language acoustically based on a well-defined basic phonetic study. The compact model is accomplished in the present invention by reducing the set of phonemes. Thus, the creation of a minimized set of phonemes
3 CA 02384518 2002-04-16 p~~CA n 2 ~ p o helps in reducing memory consumption, hence a fast execution of word editing.
Table
Table
4 represents the minimized set used in the dictation system. More specifically, Arabic words, provided as examples in Tables 1 and 4, illustrate that in the instance of gemination, only one grapheme (and not a doubled one) is used, while it is still doubled phonemically. It is also clear in the case of vowels; that is, while there are almost six degrees of vowels in table 1, and in table 4 there are only three. Hence, the difference in pronunciation is not taken into account in the written text. Accordingly, the present invention provides for a set of phonemes to be used by Arabic dictation software capable of automatic speech recognition.
BRIEF DESCRIPTION OF THE DRAWINGS
Fgure 1 illustrates the method associated with the preferred embodiment of the present invention for determining a compact model to transcribe the Arabic language.
acoustically based on a well-defined basic phonetic study.
Figure 2 illustrates in further detail the data extraction step of Figure 1.
Figure 3 illustrates the composition of the maximal set described in the method of Figure 1.
Figure 4 illustrates the various kinds of phonemes.
DESCRIP'T~N OF THE PREFERRED EMBODIMENTS
White this invention is illustrated and described in a preferred embodiment, the invention may be produced in many different configurations, forms and materials.
There is depicted in the drawings, and will herein be described in detail, a preferred embodiment of the invention, with the understanding that the present disclosure is to ~~~,~~"f~,~
be considered as an exemplification of the principles of the invention and the associated functional specifications for its construction and is not intended to limit the invention to the embodiment illustrated. Those skilled in the art will envision many other possible variations within the scope of the present invention.
One important initial step involved in the development of automatic speech recognition (ASR) software is a "basic phonetic study". A general description of such a study is starts with identifying a language on which a basic phonetic study needs to be pertormed, any material related to the phonology and phonetics of the identified language is collected (or alternatively extracted from a database over a network). This provides for an overview of the phonetic structure of the identified language.
Furthermore, technological problems and transcription problems associated with the language are identified. For example, literature in Arabic phonetics uses the terms "emphatic ; "pharyngealized", and "velarized ; which exhibit clear differences that mark their uniqueness. Addtionally, it is necessary to interpret the symbols in the literature and find a mapping to a single and more recent phonetic alphabet based on feature description rather than symbol shapes.
It should be noted that the International Phonetic Alphabet (IPA) was used in conjunction with this invention. The IPA, as defined by the International Phonetic Association (http://www.arts.gla.ac.uk/IPA/ipa.html) is a standard set of symbols for transcribing the sounds of spoken languages. The above mentioned website provides for a full chart of IPA symbols as reproduced below. Furthermore, charts for i. '' i, ~L~~:~i.
~ ~. :.~ ~, ~ s~
consonants, vowels, tones and accents, suprasegmentals, diacritics and other symbols are also provided. The last version of the IPA dates to 1993, as shown below:
PCTICA ~? 2 / p 0 5 3 0 THE INTERNATlOIaiAL P~-10NETdC AI:F'HAB~~T (wised to 1~3) ec~r~c~r~~ rrs orr~c~ .,~~
' EiWia1l,~iodewt. p,~,~ . Y~r - Z)rdr Ph~w1 . ~d Podalreol~ : Paiwl.-b ~ d . .a c - ~ '~g ~~ .
- d : ~C :,'~
.~ .
' ~ n ' ~:
n n ~
B r ',~ :'', r ~:, ' . .
v'_ :
.;~.:f 8 ~ z v : , , y ~ .-. ~ s. j, y , ~ . , ~, ~ y a x :'~ ~ ~'~..
...';
3 ...
t,,~,, 1 ~
. .
.
,~
i:
-~""oa:,tar,~,....da.~.~~~.o:.war.a.a.:.~,.d.rmarie.~i~~.,._ , _ ,.
.
.
C1J1N1"5a~0~$Mt3~NtG~ St7PR~SF~I~t~'AI:B
t.
' oia~ ' ' : W
. s~.
' ~
' W.
~ a~e~t~onoe..
~.a~,~,~.~~~b.
~
a~~
,"
' ~~
~
'~S
' s.
s ,. .
"~s: , . o .
~ j _l~na~at~r ~
p' :.~....
1 .,..~~
I. ~o~r~:
.~p;o"~.,a~i ~
. l~'1H'~
~ '~' pat~t ~
~~:. ...~:.~, ,r ,.
~
~~~b,~.
w P~trashiiol~ ~~
- ~
~ ' v~~!r ~
- .-i,ordi~
~'. ~
'~alrr .
~ ' . ~
' ~--y Low y ~i $' , .
.
.
.
~
~~
~. ~, ~ (boa . daha~t) T.
, ~
~ioW
til ..
Clo~ ~C$
~ ~
1'w. ..~al~~f~~~ar~as~e~bo1!~iba,~l:~:
'~' .'.
~~' .
~
' ' ~'.
~~
~ r~
~ ~
".
D~rl .
Y~
~
p a , w ~
b v Yes.
~, ~:,~, :, .$
.
h ~~
.,d~
~:i~.,~,~a ~.
.
.~
_v.'r"'rt ~
~.
~oa.a .
.'~
~
s~..ra~a :
t''' l~~
~
:~..a~:~e ' t.~.
~' ~,rt . ~..
(~-~'- (~
~ ' D j $ ~aru~.a .
' .
..ae~
' , ~
. . ~
. .
.
,.a ar. .t .~. :
~n d '~''d at , . y..
.
.
",~..
.
.
csa ~
~a.~,..,.~.
d' ~:.:~
~:
s ::'.~
~i ~~
J~ ..
Ydodw~'~::A'~ ~
', i~' :
~
~
~ .
' ~,e :~ur~r.~:a.~ .
, ~V
Y~.u r.p~r~aiawc .~, .X~aaa~s:~~.~
~..~..~n~:.~r ' y t~ .
.a~.;~.~w.,~a:..~ n.~::a .
~
~~
,~a.~:.~.~d . ' $ , :Vciaaa~~d~lAird~ ~:
~ .
~,;',( *a~a~~it~adior4 .
~~ ft, YOiori~ , '~aiisl~r T
~~1 ~
: ~
~"1 Mlwa~d Tc~ae'~oat , ~
:
.
. ~
~Pd~l r .~ ~' .. R~aekdTa~ue!o~
' ~
.
~~~'e~~f, r) 2 / t~ ~ 5 ~ p In the present invention, all units regarding the literature of the language in question are collected (or alternatively extracted via a database). Next, all unwanted elements are removed. This compilation allows one to establish feature sets required to describe each and all sounds of the language, and describe accurately each phonological or phonetic unit associated with the language. After the feature set and unit transcription, a representational symbol of the transcription alphabet is selected.
Subsequently, a structured table is constructed with the following information: i) all phonemes of the language, ii) all allophones of the language and their relation to the phonemes, iii) a preliminary set of rules governing the selection of allophones, iv) a set of examples, and v) the most common representation of the sounds using Roman letters.
Fgure 1 illustrates the method 100 associated with the preferred embodiment of the present invention for determining a compact model to transcribe the Arabic language acoustically (based on a well-defined basic phonetic study). First, a language for which a compact model is to be developed is identified 102. Next, information regarding the identified language is extracted or collected 104. Data extraction can be accomplished via a variety means including, but not limited to: extracting data regarding the Arabic language via a network (such as the Internet, Local Area Network (LAN), Wide Area Network (WAN) or database (local or remote). Next, from the extracted data, a list is created where the phonological and phonetic units are defined 106. As a next step, the variations in the Arabic language are identified 108.
For example, variations in classical Arabic, Modern Standard Arabic (MSA), and colloquial r~i . it~~l~~fil:
Arabic are identified. Next, a maximal set is created that contains all phonemes, allophones, and transliteration symbols associated with the Arabic language 110.
Transliteration refers to the process of representing or spelling a word (in a first language) in the characters of another alphabet (second language). Lastly, the maximal set is reduced 112 to provide for a compact set to transcribe the Arabic language acoustically. The details of the reduction step are explained in detail in the following sections.
The data extraction step of Figure 1 (102) is illustrated in further detail in Figure 2. With the extracted data, terminological problems are identified 202.
Certain terms that have been used by several phonological linguists in their attempt to defiine and describe the nature of various Arabic sounds have proved invalid; i.e. whereas few linguists may include phonemes like /F7/,/R7/, and /X/ into the category of Emphatics , others may include them in the category of pharyngeals. As a result of this non-final consensus, the most appropriate category depending upon their influence on the neighboring vowels was selected. Next, transcription problems associated with the language in question (e.g., Arabic) are identified 204. In contrast to what the IPA
exhibits in using special symbols (ASCII characters) which might cause technical problems if used in the present system; the transcription set was limited to include the characters which can be typed easily on the keyboard, Furthermore, phonological and phonetic units were extracted or collected 206 and a feature set was established based on this information 208. Next, a representative symbol for the transcription alphabet is selected 210 and a structured source is built 212. Our structured source consists of pcTicA ~z~ 00530 Phonemes, which are divided into three main units: Consonants, Vowels and Semi-Vowels. The unit "Consonants" includes a variety of Allophones and Geminations.
Aallophones may have their own gemination variety. The unit "dowels"has a variety of allophones only, while the unit "semi-vowel"has just gemination variety. The features of these units are determined according to three conditions: Place, manner of articulation in addition to the nature of the sound being voiced or voiceless.
Figure 3 illustrates the composition of the maximal set described in step 1i0 of Figure 1. Maximal set 300 comprises (but is not limited to): phonemes 302, allophones 304, a set of rules governing the selection of allophones 306, a set of examples 308, and the transliteration symbols 3i0. It should be noted that although the preferred language of this application is Arabic, one skilled in the art could extend the present invention to cover other similar languages. A detailed description of the Arabic phonetic study as per the present invention is given below.
ARABIC BASIC PHONETIC STUDY
When starting the research on basic a phonetic study for Arabic language, certain points concerning the nature of the language need to be considered. In other words, the characteristics of Arabic language at different levels (graphemic, morphological, and phonological) need to be considered. To do so, different forms of the Arabic language that can be used as an input for the text to~ speech (ITS) system need to be identified.
ARABIC LANGUAGE VARIETIES
We could distinguish in the Arabic language at least three varieties:
~...kr ., , Classic Arabic 1) Language of the Holy Qur'an; highly codified since early Islamic period 2) Used nowadays only in religious sermons or speeches Modern Standard Arabic (M54) 1) "Standard": Highly codified language, grammatically identical to classical Arabic, although case ending is not usually pronounced.
2) "Modern": Lexically adapted to modem times (e.g., lexicon innovations, loan words).
Col%guial Arabic 1) Arabic dialects: Natively teamed varieties that are used in informal situations and in the everyday communication of a geographically defined community.
Since the input for the TTS system is text, it is clear that the target language for TTS should be modem standard Arabic (MSA).
Arabic Language Has Distinctive Features Arabic letters need to be transliterated, in other words, they need to be represented by Roman alphabets in such a way that there is a one-to-one mapping between the two character systems. There is a need to not only transliterate characters, but diacritics also. Therefore, Arabic distinctive phonetic groups were created. For example, as illustrated in Fgure 4:
i) Pharyngeal phonemes like / t%K/ , / D%K/ , and / d%K / were created.
2) Emphatic phonemes like / F7/ , /R7/ , and /X/ were created.
Furthermore, Arabic language has various kinds of allophones:
ii 1!~C f.~~~f,a 1 ~7 ~ ~ ~
1) The gemination of all consonants.
Z) Normal allophones, like pharyngealized allophones of certain consonants, and the varieties of vowels.
Fnally, Arabic language has a more distinctive syllabification and lexical stress system than any other language. As a result of the basic phonetics associates with the Arabic language, a maximal set is created, that contains all phonemes, the allophones of the language, a preliminary set of rules governing the selection of allophones, a set of examples, and a transliteration symbols.
Reduction Of Maximal Set The reduction of maximal set for the TTS and the ASR phonetic sets are described in details below:
34 Phonemes 34 Phonemes Maximal 14 Allophone Deleted Minimize Setfor d setfor Geminadon T7~' Phonetic Set In the phonetic set for the TTS system, all the phonemes and allophones with which any given text message can be conveyed is found. For example, i) all the allophones for the vowels are identified; ii) allophones that represent any borrowed word in Arabic are identified, and iii) in the case of gemination, add symbols to PCTICA ~a2/ 04530 represent the phoneme when it is geminated. Thus, geminated phonemes represented by doubling the original symbol, are represented by a new symbol.
ASR Set For speech recognition, when the system recognizes the three varieties of long fatha, all of them are converted to Aiif for example, in words like /nE.OIm/, /nl.qId/
and /n2.6Ig/; the vowels E, 1, and 2 are represented by the grapheme ~. While three varieties for the short fatha will be converted into the diacritic fatha, for example, in words like /ge.le.se/, /fa:qad/ and /qAws/; the vowels e, a, and A are not represented in orthography. Both the geminated and non-geminated consonant will be represented by the same grapheme.
Only few allophones are added to recognize certain pronunciation varieties due to the country that the speech recognition system is developed for. Thus, in conclusion, the automatic speech recognition (ASR) set is less in number than the text to speech ('1'fS) set, thereby reducing the memory consumption in the resident computer system enabling easier storage of the compact set of phonetics.
The above enhancements for a compact model to transcribe the Arabic language acoustically based on a well-defined basic phonetic study are implemented in various computing environments. For example, the present invention may be implemented on a conventional computing equipment, a mufti-nodal system (e.g. LAN) or networking system (e.g. Internet, VWVW, wireless web). All programming and data related thereto are stored in computer memory, static or dynamic, and may be retrieved by the user in any of: conventional computer storage, display (i.e. CRT) and/or hardcopy (i.e. printed) ~~~'A~.f~. ~ 2 ~ ~ ~ 5 ~ ~
formats. The programming of the present invention may be implemented by one of skill in the art of automatic speech recognition (ASR).
A system and method has been shown in the above embodiments for the effective implementation of an Expanding Dictation Vocabulary. While various preferred embodiments have been shown and described, it will be understood that there is no intent to limit the invention by such disclosure, but rather, it is intended to cover all modifications and alternate constructions falling within the spirit and scope of the invention, as defined in the appended claims. For example, the present invention should not be limited by software/program, computing environment, or specific computing hardware.
BRIEF DESCRIPTION OF THE DRAWINGS
Fgure 1 illustrates the method associated with the preferred embodiment of the present invention for determining a compact model to transcribe the Arabic language.
acoustically based on a well-defined basic phonetic study.
Figure 2 illustrates in further detail the data extraction step of Figure 1.
Figure 3 illustrates the composition of the maximal set described in the method of Figure 1.
Figure 4 illustrates the various kinds of phonemes.
DESCRIP'T~N OF THE PREFERRED EMBODIMENTS
White this invention is illustrated and described in a preferred embodiment, the invention may be produced in many different configurations, forms and materials.
There is depicted in the drawings, and will herein be described in detail, a preferred embodiment of the invention, with the understanding that the present disclosure is to ~~~,~~"f~,~
be considered as an exemplification of the principles of the invention and the associated functional specifications for its construction and is not intended to limit the invention to the embodiment illustrated. Those skilled in the art will envision many other possible variations within the scope of the present invention.
One important initial step involved in the development of automatic speech recognition (ASR) software is a "basic phonetic study". A general description of such a study is starts with identifying a language on which a basic phonetic study needs to be pertormed, any material related to the phonology and phonetics of the identified language is collected (or alternatively extracted from a database over a network). This provides for an overview of the phonetic structure of the identified language.
Furthermore, technological problems and transcription problems associated with the language are identified. For example, literature in Arabic phonetics uses the terms "emphatic ; "pharyngealized", and "velarized ; which exhibit clear differences that mark their uniqueness. Addtionally, it is necessary to interpret the symbols in the literature and find a mapping to a single and more recent phonetic alphabet based on feature description rather than symbol shapes.
It should be noted that the International Phonetic Alphabet (IPA) was used in conjunction with this invention. The IPA, as defined by the International Phonetic Association (http://www.arts.gla.ac.uk/IPA/ipa.html) is a standard set of symbols for transcribing the sounds of spoken languages. The above mentioned website provides for a full chart of IPA symbols as reproduced below. Furthermore, charts for i. '' i, ~L~~:~i.
~ ~. :.~ ~, ~ s~
consonants, vowels, tones and accents, suprasegmentals, diacritics and other symbols are also provided. The last version of the IPA dates to 1993, as shown below:
PCTICA ~? 2 / p 0 5 3 0 THE INTERNATlOIaiAL P~-10NETdC AI:F'HAB~~T (wised to 1~3) ec~r~c~r~~ rrs orr~c~ .,~~
' EiWia1l,~iodewt. p,~,~ . Y~r - Z)rdr Ph~w1 . ~d Podalreol~ : Paiwl.-b ~ d . .a c - ~ '~g ~~ .
- d : ~C :,'~
.~ .
' ~ n ' ~:
n n ~
B r ',~ :'', r ~:, ' . .
v'_ :
.;~.:f 8 ~ z v : , , y ~ .-. ~ s. j, y , ~ . , ~, ~ y a x :'~ ~ ~'~..
...';
3 ...
t,,~,, 1 ~
. .
.
,~
i:
-~""oa:,tar,~,....da.~.~~~.o:.war.a.a.:.~,.d.rmarie.~i~~.,._ , _ ,.
.
.
C1J1N1"5a~0~$Mt3~NtG~ St7PR~SF~I~t~'AI:B
t.
' oia~ ' ' : W
. s~.
' ~
' W.
~ a~e~t~onoe..
~.a~,~,~.~~~b.
~
a~~
,"
' ~~
~
'~S
' s.
s ,. .
"~s: , . o .
~ j _l~na~at~r ~
p' :.~....
1 .,..~~
I. ~o~r~:
.~p;o"~.,a~i ~
. l~'1H'~
~ '~' pat~t ~
~~:. ...~:.~, ,r ,.
~
~~~b,~.
w P~trashiiol~ ~~
- ~
~ ' v~~!r ~
- .-i,ordi~
~'. ~
'~alrr .
~ ' . ~
' ~--y Low y ~i $' , .
.
.
.
~
~~
~. ~, ~ (boa . daha~t) T.
, ~
~ioW
til ..
Clo~ ~C$
~ ~
1'w. ..~al~~f~~~ar~as~e~bo1!~iba,~l:~:
'~' .'.
~~' .
~
' ' ~'.
~~
~ r~
~ ~
".
D~rl .
Y~
~
p a , w ~
b v Yes.
~, ~:,~, :, .$
.
h ~~
.,d~
~:i~.,~,~a ~.
.
.~
_v.'r"'rt ~
~.
~oa.a .
.'~
~
s~..ra~a :
t''' l~~
~
:~..a~:~e ' t.~.
~' ~,rt . ~..
(~-~'- (~
~ ' D j $ ~aru~.a .
' .
..ae~
' , ~
. . ~
. .
.
,.a ar. .t .~. :
~n d '~''d at , . y..
.
.
",~..
.
.
csa ~
~a.~,..,.~.
d' ~:.:~
~:
s ::'.~
~i ~~
J~ ..
Ydodw~'~::A'~ ~
', i~' :
~
~
~ .
' ~,e :~ur~r.~:a.~ .
, ~V
Y~.u r.p~r~aiawc .~, .X~aaa~s:~~.~
~..~..~n~:.~r ' y t~ .
.a~.;~.~w.,~a:..~ n.~::a .
~
~~
,~a.~:.~.~d . ' $ , :Vciaaa~~d~lAird~ ~:
~ .
~,;',( *a~a~~it~adior4 .
~~ ft, YOiori~ , '~aiisl~r T
~~1 ~
: ~
~"1 Mlwa~d Tc~ae'~oat , ~
:
.
. ~
~Pd~l r .~ ~' .. R~aekdTa~ue!o~
' ~
.
~~~'e~~f, r) 2 / t~ ~ 5 ~ p In the present invention, all units regarding the literature of the language in question are collected (or alternatively extracted via a database). Next, all unwanted elements are removed. This compilation allows one to establish feature sets required to describe each and all sounds of the language, and describe accurately each phonological or phonetic unit associated with the language. After the feature set and unit transcription, a representational symbol of the transcription alphabet is selected.
Subsequently, a structured table is constructed with the following information: i) all phonemes of the language, ii) all allophones of the language and their relation to the phonemes, iii) a preliminary set of rules governing the selection of allophones, iv) a set of examples, and v) the most common representation of the sounds using Roman letters.
Fgure 1 illustrates the method 100 associated with the preferred embodiment of the present invention for determining a compact model to transcribe the Arabic language acoustically (based on a well-defined basic phonetic study). First, a language for which a compact model is to be developed is identified 102. Next, information regarding the identified language is extracted or collected 104. Data extraction can be accomplished via a variety means including, but not limited to: extracting data regarding the Arabic language via a network (such as the Internet, Local Area Network (LAN), Wide Area Network (WAN) or database (local or remote). Next, from the extracted data, a list is created where the phonological and phonetic units are defined 106. As a next step, the variations in the Arabic language are identified 108.
For example, variations in classical Arabic, Modern Standard Arabic (MSA), and colloquial r~i . it~~l~~fil:
Arabic are identified. Next, a maximal set is created that contains all phonemes, allophones, and transliteration symbols associated with the Arabic language 110.
Transliteration refers to the process of representing or spelling a word (in a first language) in the characters of another alphabet (second language). Lastly, the maximal set is reduced 112 to provide for a compact set to transcribe the Arabic language acoustically. The details of the reduction step are explained in detail in the following sections.
The data extraction step of Figure 1 (102) is illustrated in further detail in Figure 2. With the extracted data, terminological problems are identified 202.
Certain terms that have been used by several phonological linguists in their attempt to defiine and describe the nature of various Arabic sounds have proved invalid; i.e. whereas few linguists may include phonemes like /F7/,/R7/, and /X/ into the category of Emphatics , others may include them in the category of pharyngeals. As a result of this non-final consensus, the most appropriate category depending upon their influence on the neighboring vowels was selected. Next, transcription problems associated with the language in question (e.g., Arabic) are identified 204. In contrast to what the IPA
exhibits in using special symbols (ASCII characters) which might cause technical problems if used in the present system; the transcription set was limited to include the characters which can be typed easily on the keyboard, Furthermore, phonological and phonetic units were extracted or collected 206 and a feature set was established based on this information 208. Next, a representative symbol for the transcription alphabet is selected 210 and a structured source is built 212. Our structured source consists of pcTicA ~z~ 00530 Phonemes, which are divided into three main units: Consonants, Vowels and Semi-Vowels. The unit "Consonants" includes a variety of Allophones and Geminations.
Aallophones may have their own gemination variety. The unit "dowels"has a variety of allophones only, while the unit "semi-vowel"has just gemination variety. The features of these units are determined according to three conditions: Place, manner of articulation in addition to the nature of the sound being voiced or voiceless.
Figure 3 illustrates the composition of the maximal set described in step 1i0 of Figure 1. Maximal set 300 comprises (but is not limited to): phonemes 302, allophones 304, a set of rules governing the selection of allophones 306, a set of examples 308, and the transliteration symbols 3i0. It should be noted that although the preferred language of this application is Arabic, one skilled in the art could extend the present invention to cover other similar languages. A detailed description of the Arabic phonetic study as per the present invention is given below.
ARABIC BASIC PHONETIC STUDY
When starting the research on basic a phonetic study for Arabic language, certain points concerning the nature of the language need to be considered. In other words, the characteristics of Arabic language at different levels (graphemic, morphological, and phonological) need to be considered. To do so, different forms of the Arabic language that can be used as an input for the text to~ speech (ITS) system need to be identified.
ARABIC LANGUAGE VARIETIES
We could distinguish in the Arabic language at least three varieties:
~...kr ., , Classic Arabic 1) Language of the Holy Qur'an; highly codified since early Islamic period 2) Used nowadays only in religious sermons or speeches Modern Standard Arabic (M54) 1) "Standard": Highly codified language, grammatically identical to classical Arabic, although case ending is not usually pronounced.
2) "Modern": Lexically adapted to modem times (e.g., lexicon innovations, loan words).
Col%guial Arabic 1) Arabic dialects: Natively teamed varieties that are used in informal situations and in the everyday communication of a geographically defined community.
Since the input for the TTS system is text, it is clear that the target language for TTS should be modem standard Arabic (MSA).
Arabic Language Has Distinctive Features Arabic letters need to be transliterated, in other words, they need to be represented by Roman alphabets in such a way that there is a one-to-one mapping between the two character systems. There is a need to not only transliterate characters, but diacritics also. Therefore, Arabic distinctive phonetic groups were created. For example, as illustrated in Fgure 4:
i) Pharyngeal phonemes like / t%K/ , / D%K/ , and / d%K / were created.
2) Emphatic phonemes like / F7/ , /R7/ , and /X/ were created.
Furthermore, Arabic language has various kinds of allophones:
ii 1!~C f.~~~f,a 1 ~7 ~ ~ ~
1) The gemination of all consonants.
Z) Normal allophones, like pharyngealized allophones of certain consonants, and the varieties of vowels.
Fnally, Arabic language has a more distinctive syllabification and lexical stress system than any other language. As a result of the basic phonetics associates with the Arabic language, a maximal set is created, that contains all phonemes, the allophones of the language, a preliminary set of rules governing the selection of allophones, a set of examples, and a transliteration symbols.
Reduction Of Maximal Set The reduction of maximal set for the TTS and the ASR phonetic sets are described in details below:
34 Phonemes 34 Phonemes Maximal 14 Allophone Deleted Minimize Setfor d setfor Geminadon T7~' Phonetic Set In the phonetic set for the TTS system, all the phonemes and allophones with which any given text message can be conveyed is found. For example, i) all the allophones for the vowels are identified; ii) allophones that represent any borrowed word in Arabic are identified, and iii) in the case of gemination, add symbols to PCTICA ~a2/ 04530 represent the phoneme when it is geminated. Thus, geminated phonemes represented by doubling the original symbol, are represented by a new symbol.
ASR Set For speech recognition, when the system recognizes the three varieties of long fatha, all of them are converted to Aiif for example, in words like /nE.OIm/, /nl.qId/
and /n2.6Ig/; the vowels E, 1, and 2 are represented by the grapheme ~. While three varieties for the short fatha will be converted into the diacritic fatha, for example, in words like /ge.le.se/, /fa:qad/ and /qAws/; the vowels e, a, and A are not represented in orthography. Both the geminated and non-geminated consonant will be represented by the same grapheme.
Only few allophones are added to recognize certain pronunciation varieties due to the country that the speech recognition system is developed for. Thus, in conclusion, the automatic speech recognition (ASR) set is less in number than the text to speech ('1'fS) set, thereby reducing the memory consumption in the resident computer system enabling easier storage of the compact set of phonetics.
The above enhancements for a compact model to transcribe the Arabic language acoustically based on a well-defined basic phonetic study are implemented in various computing environments. For example, the present invention may be implemented on a conventional computing equipment, a mufti-nodal system (e.g. LAN) or networking system (e.g. Internet, VWVW, wireless web). All programming and data related thereto are stored in computer memory, static or dynamic, and may be retrieved by the user in any of: conventional computer storage, display (i.e. CRT) and/or hardcopy (i.e. printed) ~~~'A~.f~. ~ 2 ~ ~ ~ 5 ~ ~
formats. The programming of the present invention may be implemented by one of skill in the art of automatic speech recognition (ASR).
A system and method has been shown in the above embodiments for the effective implementation of an Expanding Dictation Vocabulary. While various preferred embodiments have been shown and described, it will be understood that there is no intent to limit the invention by such disclosure, but rather, it is intended to cover all modifications and alternate constructions falling within the spirit and scope of the invention, as defined in the appended claims. For example, the present invention should not be limited by software/program, computing environment, or specific computing hardware.
Claims (20)
1. A method for determining a compact model to transcribe a language acoustically based on well-defined basic phonetics, said method comprising:
extracting phonetic information regarding said language;
defining, based on said extracted information, phonological- and phonetic units associated with said language;
identifying variations in said language;
developing a maximal set based on said defined phonological units, phonetic units, and identified variations in said language, and reducing said maximal set to a minimal set of phonemes and allophones, thereby providing for a compact model for acoustically transcribing said language.
extracting phonetic information regarding said language;
defining, based on said extracted information, phonological- and phonetic units associated with said language;
identifying variations in said language;
developing a maximal set based on said defined phonological units, phonetic units, and identified variations in said language, and reducing said maximal set to a minimal set of phonemes and allophones, thereby providing for a compact model for acoustically transcribing said language.
2. A method for determining a compact model to transcribe a language acoustically based on well-defined basic phonetics, as per claim 1, wherein said step of extracting information further comprises:
identifying terminological problems associated with said language;
identifying transcription problems associated with said language;
extracting all phonological and phonetic units associated with said language, and selecting a representative symbol for the transcription alphabet.
identifying terminological problems associated with said language;
identifying transcription problems associated with said language;
extracting all phonological and phonetic units associated with said language, and selecting a representative symbol for the transcription alphabet.
3. A method for determining a compact model to transcribe a language acoustically based on well-defined basic phonetics, as per claim 1, wherein said maximal set comprises any of, or a combination of: phonemes, allophones, rules governing the selection of allophones, a set of examples, and transliteration symbols.
4. A method for determining a compact modes to transcribe a language acoustically based on well-defined basic phonetics, as per claim 1, wherein said step of reducing said maximal set further comprises reducing an automatic speech recognition phonetic set.
5. A method for determining a compact model to transcribe a language acoustically based on well-defined basic phonetics, as per claim 4, wherein said step of reducing an automatic speech recognition phonetic set further comprises the use of diacritics, graphemes, and allophones.
6. A method for determining a compact model to transcribe a language acoustically based on well-defined basic phonetics, as per claim 1, wherein said step of reducing said maximal set further comprises reducing a text-to-speech phonetics set.
7. A method for determining a compact model to transcribe a language acoustically based on well-defined basic phonetics, as per claim 6, wherein said step of reducing an text-to-speech phonetics set is accomplished by using allophones and adding symbols representing the phoneme to be geminated.
8 A method for determining a compact model to transcribe a language acoustically based on well-defined basic phonetics, as per claim 1, wherein said transcription alphabet is in compliance with the International Phonetics Alphabet (IPA).
9. A method for determining a compact model to transcribe a language acoustically based on well-defined basic phonetics, as per claim 1, wherein said language is any of the following: modem standard Arabic (MSA), classical Arabic, or colloquial Arabic.
10. A method for determining a compact model to transcribe a language acoustically based on well-defined basic phonetics, as per claim 1, wherein said phonetic information is extracted over a network.
11. A method for determining a compact model to transcribe a language acoustically based on well-defined basic phonetics, as per claim 2, wherein said network is any of the following networks: local area networks (LAN), wide area networks (WAN), Internet, HTTP-based networks, or wireless networks.
12. A voice control system utilizing a compact model to transcribe a language acoustically based on well-defined basic phonetics, said system comprising:
a computer system;
a microphone, said microphone interfacing with said computer system, said microphone capable of receiving voice input in said language, a multimedia kit including full duplex sound card, said multimedia kit interfacing with said computer system, and said multimedia kit receiving said voice inputs from said microphone, and said computer system receiving said voice input from said multimedia kit and phonetically analyzing said voice inputs using a stored compact set of phonetic alphabets thereby enabling translation of voice-to-text based on said stored compact set of phonetic alphabets.
a computer system;
a microphone, said microphone interfacing with said computer system, said microphone capable of receiving voice input in said language, a multimedia kit including full duplex sound card, said multimedia kit interfacing with said computer system, and said multimedia kit receiving said voice inputs from said microphone, and said computer system receiving said voice input from said multimedia kit and phonetically analyzing said voice inputs using a stored compact set of phonetic alphabets thereby enabling translation of voice-to-text based on said stored compact set of phonetic alphabets.
13. A voice control system utilizing a compact model to transcribe a language acoustically based on well-defined basic phonetics, as per claim 12, wherein said multimedia kit further comprises a built-in automatic speech recognition (ASR) utility.
14. A voice control system utilizing a compact model to transcribe a language acoustically based on well-defined basic phonetics, as per claim 12, wherein said multimedia kit recognizes human voice and interprets it into corresponding actions without being speaker dependent.
15. A voice control system utilizing a compact model to transcribe a language acoustically based on well-defined basic phonetics, as per claim 14, wherein said speaker dependant includes gender or age.
16. A voice control system utilizing a compact model to transcribe a language acoustically based on well-defined basic phonetics, as per claim 12, wherein said compact set of phonetic alphabets is accomplished using diacritics, graphemes, and allophones.
17. A voice control system utilizing a compact model to transcribe a language acoustically based on well-defined basic phonetics, as per claim 12, wherein said compact set of phonetic alphabets are compliant with the International Phonetics Alphabet (IPA) standard.
18. A voice control system utilizing a compact model to transcribe a language acoustically based on well-defined basic phonetics, as per claim 12, wherein said language is any of the following: modem standard Arabic, classical Arabic, or colloquial Arabic.
19. A voice control method utilizing a compact model to transcribe a language acoustically based on well-defined basic phonetics, said method comprising:
receiving voice inputs in said language via a microphone;
phonetically analyzing said received voice inputs using a computer-based system, and said computer-based system analyzing said voice input using a stored compact set of phonetic alphabets, thereby enabling translation of voice-to-text based on said stored compact set of phonetic alphabets.
receiving voice inputs in said language via a microphone;
phonetically analyzing said received voice inputs using a computer-based system, and said computer-based system analyzing said voice input using a stored compact set of phonetic alphabets, thereby enabling translation of voice-to-text based on said stored compact set of phonetic alphabets.
20. A voice control method utilizing a compact model to transcribe a language acoustically based on well-defined basic phonetics, as per claim 19, wherein said compact set of phonetic alphabets is accomplished using diacritics, graphemes, and allophones.
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US09/835,535 US7107215B2 (en) | 2001-04-16 | 2001-04-16 | Determining a compact model to transcribe the arabic language acoustically in a well defined basic phonetic study |
US09/835,535 | 2001-04-16 |
Publications (1)
Publication Number | Publication Date |
---|---|
CA2384518A1 true CA2384518A1 (en) | 2002-10-16 |
Family
ID=25269758
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CA002384518A Abandoned CA2384518A1 (en) | 2001-04-16 | 2002-04-16 | Determining a compact model to transcribe the arabic language acoustically in a well defined basic phonetic study |
Country Status (3)
Country | Link |
---|---|
US (1) | US7107215B2 (en) |
EP (1) | EP1251490A1 (en) |
CA (1) | CA2384518A1 (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111008824A (en) * | 2019-12-13 | 2020-04-14 | 陇东学院 | Scientific research help-seeking management system based on multi-disciplinary fusion of network and geographic information technology |
Families Citing this family (52)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6801763B2 (en) * | 1997-10-29 | 2004-10-05 | Metro One Telecommunications, Inc. | Technique for effectively communicating travel directions |
US7177792B2 (en) * | 2001-05-31 | 2007-02-13 | University Of Southern California | Integer programming decoder for machine translation |
US8214196B2 (en) | 2001-07-03 | 2012-07-03 | University Of Southern California | Syntax-based statistical translation model |
EP1483687A4 (en) * | 2002-03-11 | 2008-08-06 | Univ Southern California | Named entity translation |
WO2004001623A2 (en) | 2002-03-26 | 2003-12-31 | University Of Southern California | Constructing a translation lexicon from comparable, non-parallel corpora |
US7389228B2 (en) * | 2002-12-16 | 2008-06-17 | International Business Machines Corporation | Speaker adaptation of vocabulary for speech recognition |
US7711545B2 (en) * | 2003-07-02 | 2010-05-04 | Language Weaver, Inc. | Empirical methods for splitting compound words with application to machine translation |
US8548794B2 (en) | 2003-07-02 | 2013-10-01 | University Of Southern California | Statistical noun phrase translation |
WO2005089340A2 (en) * | 2004-03-15 | 2005-09-29 | University Of Southern California | Training tree transducers |
US8296127B2 (en) * | 2004-03-23 | 2012-10-23 | University Of Southern California | Discovery of parallel text portions in comparable collections of corpora and training using comparable texts |
US8666725B2 (en) | 2004-04-16 | 2014-03-04 | University Of Southern California | Selection and use of nonstatistical translation components in a statistical machine translation framework |
US7430503B1 (en) * | 2004-08-24 | 2008-09-30 | The United States Of America As Represented By The Director, National Security Agency | Method of combining corpora to achieve consistency in phonetic labeling |
JP5452868B2 (en) | 2004-10-12 | 2014-03-26 | ユニヴァーシティー オブ サザン カリフォルニア | Training for text-to-text applications that use string-to-tree conversion for training and decoding |
US9122655B2 (en) * | 2004-11-15 | 2015-09-01 | International Business Machines Corporation | Pre-translation testing of bi-directional language display |
US8886517B2 (en) | 2005-06-17 | 2014-11-11 | Language Weaver, Inc. | Trust scoring for language translation systems |
US8676563B2 (en) | 2009-10-01 | 2014-03-18 | Language Weaver, Inc. | Providing human-generated and machine-generated trusted translations |
US7974833B2 (en) | 2005-06-21 | 2011-07-05 | Language Weaver, Inc. | Weighted system of expressing language information using a compact notation |
US7389222B1 (en) | 2005-08-02 | 2008-06-17 | Language Weaver, Inc. | Task parallelization in a text-to-text system |
US7813918B2 (en) * | 2005-08-03 | 2010-10-12 | Language Weaver, Inc. | Identifying documents which form translated pairs, within a document collection |
KR100739726B1 (en) * | 2005-08-30 | 2007-07-13 | 삼성전자주식회사 | Method and system for name matching and computer readable medium recording the method |
US7624020B2 (en) * | 2005-09-09 | 2009-11-24 | Language Weaver, Inc. | Adapter for allowing both online and offline training of a text to text system |
US10319252B2 (en) | 2005-11-09 | 2019-06-11 | Sdl Inc. | Language capability assessment and training apparatus and techniques |
KR101265263B1 (en) * | 2006-01-02 | 2013-05-16 | 삼성전자주식회사 | Method and system for name matching using phonetic sign and computer readable medium recording the method |
US8943080B2 (en) | 2006-04-07 | 2015-01-27 | University Of Southern California | Systems and methods for identifying parallel documents and sentence fragments in multilingual document collections |
US8886518B1 (en) | 2006-08-07 | 2014-11-11 | Language Weaver, Inc. | System and method for capitalizing machine translated text |
US8433556B2 (en) | 2006-11-02 | 2013-04-30 | University Of Southern California | Semi-supervised training for statistical word alignment |
US9122674B1 (en) | 2006-12-15 | 2015-09-01 | Language Weaver, Inc. | Use of annotations in statistical machine translation |
US8468149B1 (en) | 2007-01-26 | 2013-06-18 | Language Weaver, Inc. | Multi-lingual online community |
US8615389B1 (en) | 2007-03-16 | 2013-12-24 | Language Weaver, Inc. | Generation and exploitation of an approximate language model |
US8831928B2 (en) | 2007-04-04 | 2014-09-09 | Language Weaver, Inc. | Customizable machine translation service |
US20080300861A1 (en) * | 2007-06-04 | 2008-12-04 | Ossama Emam | Word formation method and system |
US8825466B1 (en) | 2007-06-08 | 2014-09-02 | Language Weaver, Inc. | Modification of annotated bilingual segment pairs in syntax-based machine translation |
US8719016B1 (en) | 2009-04-07 | 2014-05-06 | Verint Americas Inc. | Speech analytics system and system and method for determining structured speech |
JP5413622B2 (en) * | 2009-04-30 | 2014-02-12 | 日本電気株式会社 | Language model creation device, language model creation method, and program |
US8990064B2 (en) | 2009-07-28 | 2015-03-24 | Language Weaver, Inc. | Translating documents based on content |
US8380486B2 (en) | 2009-10-01 | 2013-02-19 | Language Weaver, Inc. | Providing machine-generated translations and corresponding trust levels |
WO2011059800A1 (en) * | 2009-10-29 | 2011-05-19 | Gadi Benmark Markovitch | System for conditioning a child to learn any language without an accent |
US10417646B2 (en) | 2010-03-09 | 2019-09-17 | Sdl Inc. | Predicting the cost associated with translating textual content |
US8433557B2 (en) * | 2010-05-07 | 2013-04-30 | Technology Development Center, King Abdulaziz City For Science And Technology | System and method of transliterating names between different languages |
US8473280B2 (en) * | 2010-08-06 | 2013-06-25 | King Abdulaziz City for Science & Technology | System and methods for cost-effective bilingual texting |
US11003838B2 (en) | 2011-04-18 | 2021-05-11 | Sdl Inc. | Systems and methods for monitoring post translation editing |
US8694303B2 (en) | 2011-06-15 | 2014-04-08 | Language Weaver, Inc. | Systems and methods for tuning parameters in statistical machine translation |
US8886515B2 (en) | 2011-10-19 | 2014-11-11 | Language Weaver, Inc. | Systems and methods for enhancing machine translation post edit review processes |
US8942973B2 (en) | 2012-03-09 | 2015-01-27 | Language Weaver, Inc. | Content page URL translation |
US10261994B2 (en) | 2012-05-25 | 2019-04-16 | Sdl Inc. | Method and system for automatic management of reputation of translators |
US9966064B2 (en) | 2012-07-18 | 2018-05-08 | International Business Machines Corporation | Dialect-specific acoustic language modeling and speech recognition |
US9152622B2 (en) | 2012-11-26 | 2015-10-06 | Language Weaver, Inc. | Personalized machine translation via online adaptation |
US9213694B2 (en) | 2013-10-10 | 2015-12-15 | Language Weaver, Inc. | Efficient online domain adaptation |
WO2015161493A1 (en) * | 2014-04-24 | 2015-10-29 | Motorola Solutions, Inc. | Method and apparatus for enhancing alveolar trill |
US10002543B2 (en) * | 2014-11-04 | 2018-06-19 | Knotbird LLC | System and methods for transforming language into interactive elements |
US20170148341A1 (en) * | 2015-11-25 | 2017-05-25 | David A. Boulton | Methodology and system for teaching reading |
US11783813B1 (en) * | 2021-05-02 | 2023-10-10 | Abbas Rafii | Methods and systems for improving word discrimination with phonologically-trained machine learning models |
Family Cites Families (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO1995002879A1 (en) * | 1993-07-13 | 1995-01-26 | Theodore Austin Bordeaux | Multi-language speech recognition system |
US5933804A (en) * | 1997-04-10 | 1999-08-03 | Microsoft Corporation | Extensible speech recognition system that provides a user with audio feedback |
US5930754A (en) * | 1997-06-13 | 1999-07-27 | Motorola, Inc. | Method, device and article of manufacture for neural-network based orthography-phonetics transformation |
US6016470A (en) * | 1997-11-12 | 2000-01-18 | Gte Internetworking Incorporated | Rejection grammar using selected phonemes for speech recognition system |
US5953701A (en) * | 1998-01-22 | 1999-09-14 | International Business Machines Corporation | Speech recognition models combining gender-dependent and gender-independent phone states and using phonetic-context-dependence |
US6490557B1 (en) * | 1998-03-05 | 2002-12-03 | John C. Jeppesen | Method and apparatus for training an ultra-large vocabulary, continuous speech, speaker independent, automatic speech recognition system and consequential database |
US6208968B1 (en) * | 1998-12-16 | 2001-03-27 | Compaq Computer Corporation | Computer method and apparatus for text-to-speech synthesizer dictionary reduction |
US6363342B2 (en) * | 1998-12-18 | 2002-03-26 | Matsushita Electric Industrial Co., Ltd. | System for developing word-pronunciation pairs |
DE19920501A1 (en) * | 1999-05-05 | 2000-11-09 | Nokia Mobile Phones Ltd | Speech reproduction method for voice-controlled system with text-based speech synthesis has entered speech input compared with synthetic speech version of stored character chain for updating latter |
US6434521B1 (en) * | 1999-06-24 | 2002-08-13 | Speechworks International, Inc. | Automatically determining words for updating in a pronunciation dictionary in a speech recognition system |
US6738738B2 (en) * | 2000-12-23 | 2004-05-18 | Tellme Networks, Inc. | Automated transformation from American English to British English |
US7035794B2 (en) * | 2001-03-30 | 2006-04-25 | Intel Corporation | Compressing and using a concatenative speech database in text-to-speech systems |
-
2001
- 2001-04-16 US US09/835,535 patent/US7107215B2/en not_active Expired - Fee Related
-
2002
- 2002-04-16 EP EP02008591A patent/EP1251490A1/en not_active Ceased
- 2002-04-16 CA CA002384518A patent/CA2384518A1/en not_active Abandoned
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111008824A (en) * | 2019-12-13 | 2020-04-14 | 陇东学院 | Scientific research help-seeking management system based on multi-disciplinary fusion of network and geographic information technology |
CN111008824B (en) * | 2019-12-13 | 2023-08-18 | 陇东学院 | Scientific research help-seeking management system based on network and geographic information technology multidisciplinary fusion |
Also Published As
Publication number | Publication date |
---|---|
US7107215B2 (en) | 2006-09-12 |
US20030040909A1 (en) | 2003-02-27 |
EP1251490A1 (en) | 2002-10-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP1251490A1 (en) | Compact phonetic model for arabic languages recognition | |
Prahallad et al. | The IIIT-H Indic speech databases. | |
El-Imam | Phonetization of Arabic: rules and algorithms | |
US20020198715A1 (en) | Artificial language generation | |
US20070255567A1 (en) | System and method for generating a pronunciation dictionary | |
JP2001296880A (en) | Method and device to generate plural plausible pronunciation of intrinsic name | |
KR20050070073A (en) | Scalable neural network-based language identification from written text | |
Alsharhan et al. | Improved Arabic speech recognition system through the automatic generation of fine-grained phonetic transcriptions | |
Chan et al. | Development of a Cantonese-English code-mixing speech corpus. | |
US10930274B2 (en) | Personalized pronunciation hints based on user speech | |
Ashour | Major differences between Arabic and English pronunciation systems: a contrastive analysis study | |
Sarma et al. | Development and analysis of speech recognition systems for Assamese language using HTK | |
Pike | Selected writings: to commemorate the 60th birthday of Kenneth Lee Pike | |
Wells | Phonetic transcription and analysis | |
Maamouri et al. | Dialectal Arabic telephone speech corpus: Principles, tool design, and transcription conventions | |
Tseng et al. | Machine readable phonetic transcription system for Chinese dialects spoken in Taiwan | |
Keys | First language influence on the spoken English of Brazilian students of EFL | |
Soman et al. | Corpus driven malayalam text-to-speech synthesis for interactive voice response system | |
Sindran et al. | Rule-based standard Arabic Phonetization at phoneme, allophone, and syllable level | |
Marasek et al. | Multi-level annotation in SpeeCon Polish speech database | |
Dijkstra et al. | Frisian TTS, an example of bootstrapping TTS for minority languages | |
Chao-angthong et al. | Northern Thai dialect text to speech | |
Hussain | To-sound conversion for Urdu text-to-speech system | |
Zitouni et al. | OrienTel: speech-based interactive communication applications for the mediterranean and the Middle East | |
Li et al. | A grapheme to phoneme converter for standard Malay |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
EEER | Examination request | ||
FZDE | Discontinued |