WO2010062542A1 - Method for translation of a communication between languages, and associated system and computer program product - Google Patents

Method for translation of a communication between languages, and associated system and computer program product Download PDF

Info

Publication number
WO2010062542A1
WO2010062542A1 PCT/US2009/062030 US2009062030W WO2010062542A1 WO 2010062542 A1 WO2010062542 A1 WO 2010062542A1 US 2009062030 W US2009062030 W US 2009062030W WO 2010062542 A1 WO2010062542 A1 WO 2010062542A1
Authority
WO
Grant status
Application
Patent type
Prior art keywords
humanly
language
communication
perceptible
translation
Prior art date
Application number
PCT/US2009/062030
Other languages
French (fr)
Inventor
Abhijat Gupta
Kevin Anthony Wilson
Brent Thomas Ward
Original Assignee
Research Triangle Institute
Priority date (The priority date 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 date listed.)
Filing date
Publication date

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/20Handling natural language data
    • G06F17/28Processing or translating of natural language
    • G06F17/2809Data driven translation
    • G06F17/2818Statistical methods, e.g. probability models
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/20Handling natural language data
    • G06F17/28Processing or translating of natural language
    • G06F17/2809Data driven translation
    • G06F17/2845Using very large corpora, e.g. the world wide web [WWW]

Abstract

A method is provided for improving translation of a communication between humanly-perceptible languages. Such a method comprises collecting communication data in various manners from a communication available in more than one humanly- perceptible language. The communication data may be collected in various manners, and may include user-generated translations, reference translations, and translation evaluations. The collected communication data is then used to create and/or improve inter-language, intra-language, and/or other correlations, which may be applicable to translation of subsequent communications between humanly-perceptible languages. Associated methods, systems and computer program products are also provided.

Description

METHOD FOR TRANSLATION OF A COMMUNICATION BETWEEN LANGUAGES, AND ASSOCIATED SYSTEM AND COMPUTER PROGRAM

PRODUCT

BACKGROUND OF THE INVENTION

Field of the Invention

[0001] Embodiments of the present invention are generally directed to language translations and, more particularly, to a method for translation of a communication between humanly-perceptible languages, and associated methods, systems and computer program products.

Description of Related Art

[0002] As the global economy continues to grow, interaction among entities speaking, comprehending, or otherwise understanding different languages continues to grow as well. For example, documents and/or communications (e.g., email, letters, etc.) may be transmitted between such entities, wherein such documents and/or communications must be either translated by a human-translator (typically employed by a translation service) or a machine (i.e., a computer device) before reaching an intended recipient. In general, human-translators provide more accurate and higher quality translations than their machine counterparts. However, translations performed by human-translators may often be slower and more costly than those performed by the machine translators. Also, translations may differ between individual translators due to, for example, skill level, personal and/or environmental characteristics, demographics, etc. Because of these factors, machine translations are typically considered the preferred translation method, especially when transmitting electronic communications such as email. However, there is a significant need for improved quality of such machine translations, wherein words/phrases may be, for example, mistranslated, translated out of context, translated according to "mechanistic" rules, or simply not translated by the machine translators. [0003] In some instances, machine translators may employ rule-based translation schemes operating on rule-based translation dictionaries that perform translations based on a one-to-one matching of sentences, phrases, and/or words. One limitation of implementing such "mechanistic" rule-based translation dictionaries is that words may have different translations depending on the context in which they are used, which may not be accounted for by rule-based translation schemes. To account for such context issues that may limit the effectiveness of rule-based translation schemes, machine translations may instead employ a statistical translation scheme based on statistical translation models/engines. However, such statistical translation models/engines typically lack sufficient quantities of data to build/enhance/train the translation models/engines. As such, entities interacting in the global economy may turn to the less consistent, less cost-efficient and less time-efficient method of employing a human-translator to appropriately translate a communication.

[0004] Thus, there exists a need for methods, systems, and/or computer program products which provide an improved and/or customized translation of a communication, while continuously improving the quality of a translation, in a consistent, time-efficient and cost-efficient manner.

BRIEF SUMMARY OF THE INVENTION [0005] The above and other needs are met by embodiments of the present invention which, according to one aspect, provides a method for improving translation of a communication between a first humanly-perceptible language and a second humanly- perceptible language. Such a method comprises collecting communication data from a first communication in a first humanly-perceptible language and a second humanly- perceptible language. A portion of the first communication in the first humanly- perceptible language is compared with a corresponding portion of the first communication in the second humanly-perceptible language so as to determine an inter-language correlation between the first and second humanly-perceptible languages. The inter-language correlation is applicable to translate a portion of a second communication between the first humanly-perceptible language and the second humanly-perceptible language.

[0006] Another aspect of the invention provides a method for improving translation of a communication between a first humanly-perceptible language and a second humanly-perceptible language. Such a method comprises providing a statement in a first humanly-perceptible language to an audience, and soliciting a translation of the statement from the audience in a second humanly-perceptible language and receiving the translation. A portion of the statement in the first humanly-perceptible language is compared with a corresponding portion of the translation in the second humanly- perceptible language so as to determine a correlation between the first and second humanly-perceptible languages. The correlation is applicable to translate a portion of a communication between the first humanly-perceptible language and the second humanly-perceptible language.

[0007] Still another aspect of the present invention provides a method for improving translation of a communication between a first humanly-perceptible language and a second humanly-perceptible language. Such a method comprises soliciting one of a translation of a first communication in a first humanly-perceptible language between the first humanly-perceptible language and a second humanly- perceptible language, and an evaluation of the first communication in the second humanly-perceptible language, from a translator and receiving the one of the translation and the evaluation. A portion of the first communication in the first humanly- perceptible language is compared with one of a corresponding portion of the first communication translated into the second humanly-perceptible language by the translator and a corresponding portion of the first communication in the second humanly-perceptible language evaluated by the translator so as to determine an inter- language correlation between the first and second humanly-perceptible languages. The inter-language correlation is applicable to translate a portion of a second communication between the first humanly-perceptible language and the second humanly-perceptible language. [0008] Another aspect of the present invention provides a method for improving translation of a communication between a first humanly-perceptible language and a second humanly-perceptible language. Such a method comprises collecting communication data from a first communication in a first humanly-perceptible language and a second humanly-perceptible language and associating the communication data with a general domain. The communication data is evaluated with respect to a domain threshold criteria. Above-threshold communication data is associated with a specific domain. A portion of the above-threshold communication data in the first humanly-perceptible language is compared with a corresponding portion of the above-threshold communication data in the second humanly-perceptible language so as to determine a specific domain correlation between the first and second humanly-perceptible languages. The specific domain correlation is applicable to translate a portion of a second communication, associated with the specific domain, between the first humanly-perceptible language and the second humanly-perceptible language. [0009] Yet another aspect of the present invention provides a method for improving translation of a communication between a first humanly-perceptible language and a second humanly-perceptible language, the first and second humanly-perceptible language being associated with a common character script. Such a method comprises receiving a text communication comprising the common character script arranged according to a first humanly-perceptible language, and correlating the common character script of the text communication with communication data in the first humanly-perceptible language. The correlated communication data in the first humanly-perceptible language is compared with corresponding communication data in a second humanly-perceptible language so as to determine a correlation between the common character script of the text communication arranged according to the first humanly-perceptible language and the second humanly-perceptible language. The correlation is applicable to translate a portion of a second text communication comprising the common character script arranged according to the first humanly- perceptible language and the second humanly-perceptible language.

[0010] Embodiments of the present invention thus provide significant advantages as further detailed herein.

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWING(S) [0011] Having thus described the invention in general terms, reference will now be made to the accompanying drawings, which are not necessarily drawn to scale, and wherein:

[0012] FIG. 1 is a schematic of an exemplary computer device capable of translating a communication from a first human-perceptible language to at least a second human-perceptible language;

[0013] FIG. 2 is a schematic of an electronic communication translation system with an electronic communication being received by an appropriately-configured server device;

[0014] FIG. 3 is a chart illustrating a method for improving translation of a communication from a first humanly-perceptible language to a second humanly- perceptible language, in accordance with one embodiment of the present invention; [0015] FIG. 4 is a chart illustrating another method for improving translation of a communication from a first humanly-perceptible language to a second humanly- perceptible language, in accordance with one embodiment of the present invention; [0016] FIG. 5 is a chart illustrating yet another method for improving translation of a communication from a first humanly-perceptible language to a second humanly- perceptible language, in accordance with one embodiment of the present invention; [0017] FIG. 6 is a chart illustrating still yet another method for improving translation of a communication from a first humanly-perceptible language to a second humanly-perceptible language, in accordance with one embodiment of the present invention;

[0018] FIG. 7 is a chart illustrating still another method for improving translation of a communication from a first humanly-perceptible language to a second humanly- perceptible language, in accordance with one embodiment of the present invention; [0019] FIG. 8 is a chart illustrating a method for customizing translation of a communication between a first humanly-perceptible language and a second humanly- perceptible language, in accordance with one embodiment of the present invention; [0020] FIG. 9 is a chart illustrating a method for facilitating customized translation of a communication between a first humanly-perceptible language and a second humanly-perceptible language, in accordance with one embodiment of the present invention;

[0021] FIG. 10 is a chart illustrating another method for facilitating customized translation of a communication between a first humanly-perceptible language and a second humanly-perceptible language, in accordance with one embodiment of the present invention; and

[0022] FIG. 11 is a chart illustrating a method for facilitating text communication, in accordance with one embodiment of the present invention.

DETAILED DESCRIPTION OF THE INVENTION

[0023] The present invention now will be described more fully hereinafter with reference to the accompanying drawings, in which some, but not all embodiments of the inventions are shown. Indeed, these inventions may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this disclosure will satisfy applicable legal requirements. Like numbers refer to like elements throughout. [0024] Embodiments of the present invention are directed to methods, systems, and computer program products capable of improving and/or customizing translations of communications such as electronic documents, electronic communications (e.g., e- mails, instant messages, text messages, or the like), and other forms of interaction (text or oral). For example, an email communication transmitted between entities communicating in different languages, or having different language preferences, may need to be translated before the email communication reaches the intended recipient. In such an instance, the sender of the email communication may require a certain confidence that the email communication has been accurately translated by a machine translator with an appropriate level of quality.

[0025] In general, the methods, systems, and computer program products disclosed herein provide different mechanisms for improving and/or customizing translation of a communication by collecting translation data and/or other additional translation material on a continuous, semi-continuous, iterative or batchwise basis, and incorporating such translation data / additional translation material into a translation engine/model which, in some instances, may be a statistical translation model. In such a manner, the translation data / additional translation material provides the engine/model with a further improved basis (i.e., further refined statistics) for application to future translations of communications. Such translation data / additional translation material may take many different forms, including, but not limited to, user- generated feedback, reference translations and/or dictionary editing. User-generated feedback can include information from surveys, as well as comments reflecting translation errors (or providing translation evaluations) and corrected translations. Reference translations are translations that have been validated as accurate and/or preferred. Such, reference translations may be useful, for example, in building or enhancing statistical translation models/engines and, in some instances, may be weighted more heavily than other translation data / additional translation material to reflect this relatively higher value. Dictionary editing is a general term that can include both modifications to a translation dictionary/database as well as the creation of the dictionary/database itself. Further, such editing can be done manually by humans, or automatically, through the use of, for example, scoring and weighting algorithms. Because some translation engines may need to be rebuilt / restructured in order to incorporate such new translation material, and because such rebuilding / restructuring can take time, edits / updates / modifications to such dictionaries may more effectively be done in a batch processing mode.

[0026] As previously discussed, machine translations, which facilitate translation through non-human translators (i.e., translations are completed by a computer device or an otherwise automated device) may often be the preferred method for translating documents and/or communications due to, for example, time and cost restraints as well as overall efficiency. Such machine translations may be generally limited in their application due to, for example, a lack of translation data for supporting the translation engine and/or lack of a mechanism for receiving feedback for improving the translation engine employed by the machine. As such, low-quality, incomplete, or otherwise insufficient translations may result from such "conventional" machine translations. Further, words/phrases translated by a machine translator may be, for example, taken out of context when translated or, in other instances, not translated at all because the word/phrase is not recognized or otherwise understood by the translation scheme/engine/model. In this regard, the methods, systems, and computer program products disclosed herein are generally directed toward improving and/or customizing translations of communications by opportunistically collecting/extracting translation data so as to provide a continually improved basis on which to build/train/enhance the translation model/engine associated with a machine translator. Such machine translations may be accomplished by, for example, a computer device executing an appropriate computer program product. For purposes of this discussion, it is recognized and understood that the term 'computer device' includes, but is not limited to, desktop and laptop computers, as well as cellular phones, personal digital assistants (PDA), and other electronic devices, both portable and non-portable, having processing capabilities.

[0027] As an example, it may be desirable to improve a machine translation such that translated documents can be transmitted between professional entities (e.g., law firms) with a high level of confidence that the document has been translated in a manner sufficient to accurately convey the information contained therein. In other instances, it may be desirable to customize a machine translation to account for a user's native dialect such that a personalized translation is provided to that user. In any instance, improved and/or customized translations that provide higher quality and more accurate results, particularly when implemented in a machine translator, are needed for a fast-paced and global economy unwilling to wait and/or compensate for a human- translation of a document or other communication.

[0028] Such machine translations may be accomplished, as shown, for example, in FIG. 1, with respect to a computer device 10, wherein the computer device 10 may be, for example, a laptop/desktop computer, a cellular phone, a PDA, or other electronic device having processing capabilities. In this regard, a communication in a first human-perceptible language may be inputted into the computer device 10 with an input device 12, such as, for example, a keyboard, an auditory input device (e.g., a microphone) in communication with voice-recognition software, or any other suitable input device. In some instances, the communication may be received by the computer device 10 from another computer device 10. In any instance, the communication may be entered into or otherwise received by a translation scheme/engine/model embodied as a computer program product 14 associated with and executable by the computer device 10. [0029] In some instances, the translation scheme/program 14 may be locally stored and executed on the computer device 10, while in other instances the translation scheme/program 14 may be hosted or otherwise associated with a website accessible via a network, such as the Internet. In such instances, the communication may be submitted in the first human-perceptible language for translation at the website and, in return, a translation thereof in a second human-perceptible language is subsequently provided. In any instance, the translation scheme/program 14 may employ a translation model/engine, such as a statistical translation model, configured to translate and/or otherwise manipulate the communication between the first humanly-perceptible language and a second humanly-perceptible language. That is, in one aspect, the translation scheme/program may utilize a statistical machine translation (SMT) system to translate the content of the communication, though one skilled in the art will appreciate that other translation systems/schemas may also be implemented, whether statistically based or not. The translated communication may then be outputted to or via an output device 16, such as, for example, a computer screen interface, a printer device, an audio device, a tactile device or any other appropriate output device.

[0030] In some instances, the translation of the communication may occur before, during, or after transmission of an electronic communication (email, text message, instant message, etc.) or an electronic document. In such instances, the translation program/scheme 14 may be associated with an email platform/computer program product. Transmission of the electronic communication may be typically accomplished, as shown, for example, in FIG. 2, with respect to an electronic communication system 1 wherein one or more first computer devices 10 are capable of communicating through a first server device 20, over a communications network 50 and through a second server device 25, with one or more second computer devices 75. The translation process may be seamless and invisible to both user and recipient, possibly with the exception of the automatic inclusion of the source-language communication ("original communication/document" in the first humanly perceptible language) in the message received by the recipient (whereby the recipient may receive both the original communication/document and the translated communication/document) and the destination-language text ("translated communication/document" in the second humanly-perceptible language) received by the sender (whereby the sender may receive the translated communication/document for confirmation of the extent and nature of the translation). For purposes of this discussion herein, it is recognized and understood that the term 'computer network' includes, but is not limited to, computers connected via networks such as the Internet and similar protocols, as well as computers, cellular phones and other electronic devices connected via wireless and/or wireline networks. It is further recognized and understood that the term 'wireless networks' can include cellular, wideband, satellite and any other system using electromagnetic radiation for the purpose of communication.

[0031] In any instance, in order to facilitate the improved/customized translation process disclosed herein, embodiments of the present invention may be based on a translation scheme built, enhanced, and/or trained by incorporation and recognition of a relatively large amount of translation data received in a continuous, semi-continuous and/or batchwise manner. That is, a translation model/engine, such as a statistical translation model/engine, may be improved, enhanced, and/or customized by increasing the quantity of translation data incorporated therein on which the engine can base future translations. According to embodiments of the present invention, such translation data may be, for example, personal in nature (i.e., based on demographics of a person seeking a translation), or based in the collection of translation information from text, documents, or other content translated into different humanly-perceptible languages by various sources to create inter-language or other correlations that may be used for translations of communications. [0032] Previously, some translation models/engines were derived from text content obtained from proceedings of international political entities such as, for example, the European Parliament and the United Nations, due to the availability of quantities of corresponding information in multiple languages. However, the translation models/engines derived from such content sources may be context-specific (i.e., based on a single "domain" such as a political proceeding). That is, such translation models/engines would be derived from translated information specific to a context or domain directed to diplomatic/political issues and/or other issues particularly discussed by the European Parliament or the United Nations, while in session. To that end, such prior translation models/engines may have been adequate to translate text content produced from proceedings of the European Parliament or the United Nations. However, when applied to communications involving other contexts or "domains" (i.e., sports, military, medicine, etc.), such prior translation models/engines were not necessarily able to identify or adapt to the difference in context and, as such, may have provided less-than-adequate translations of communications in other contexts or domains.

[0033] Accordingly, some embodiments of the present invention are directed to compiling/collecting general domain content (i.e., over many domains / contexts) that may be parsed, analyzed, and applied across various specific domains to provide improved and/or customized translations in a particular domain, rather than collecting and applying translation data specific to a particular domain and applying that data to any and all domains. In one aspect of the present invention, a method for improving translation of a communication is provided, wherein the communication may be, for example, a document, an email, a text message, an instant message, text content, oral content (i.e., received via a voice recognition computer program product) or any other communicative message.

[0034] As shown in FIG. 3, such a method may comprise collecting communication data from a first communication in a first humanly-perceptible language and a second humanly-perceptible language, as represented by step 100. For example, a literary text/document/work may often be published or otherwise made available in multiple languages. As such, these literary works may provide substantial quantities of translation data that may be used, in accordance with the embodiments of the present invention, to compile translation data associated with a general context/domain that can be used to enhance/train/build translation models/engines, such as a statistical translation model. Such a method may further comprise comparing a portion of the first communication in the first humanly-perceptible language with a corresponding portion of the first communication in the second humanly-perceptible language so as to determine an inter-language or other correlation between the first and second humanly- perceptible languages, as represented by step 102. In this manner, the inter-language correlation may be applicable to translate a portion of a second communication between the first humanly-perceptible language and the second humanly-perceptible language. In some instances, the inter-language correlation may be, for example, a statistical translation model of a statistical machine translator. [0035] In one exemplary embodiment, a literary work published or otherwise available in a first humanly-perceptible language and a second humanly-perceptible language may be scanned using an appropriate scanner device such that the text content thereof may be converted into appropriate code usable by a computer device through, for example, optical character recognition technology, as understood by one having skill in the art. A verification/evaluation program associated with the computer device may also be used to determine the accuracy of the scanned, extracted, or otherwise converted text content. In some instances, the verification program may also be used to automatically correct any scanning errors, including, for example, those associated with the optical character recognition. The text content determined from each language version of the single literary work published in the first and second humanly- perceptible languages may be appropriately correlated using a computer algorithm or other appropriate procedure such that corresponding portions of the text content of the work in the first and second humanly-perceptible languages are capable of being compared to create an inter-language correlation. Such a correlation may involve a word-to-word correspondence, as well as, for example, a grammar-, a context-, and/or a structural correspondence.

[0036] In some instances, at least a portion of the literary work in the first humanly- perceptible language and/or the second humanly-perceptible language may be compared with a corresponding language reference model (i.e., a recognized standard or a generally-accepted dictionary) so as to determine an accuracy of the literary work in the respective humanly-perceptible language. The communication data and/or inter- language correlation may be accessible to or otherwise associated with a translation model/engine, such as a statistical translation model/engine, such that the communication data and/or inter-language correlation is applicable to a translation of a second communication, which may be subsequently submitted for translation, for example, to a computer device executing a computer program product configured to provide the translation, to an Internet website, or via a computer network, as described previously. In any instance, the translation information collected from the literary work published in the first and second humanly-perceptible languages can be converted, extracted, manipulated, and applied to enhance a translation model/engine such that subsequent translations are, for example, improved, more accurate, and generally of a higher quality. While reference has been made to a literary work, it will be understood that such a communication used for collecting the translation data and other communication data may be any type of work having content, including, but not limited to, books, journals, documents, and the like.

[0037] In another aspect of the present invention, a method for improving translation of a communication between a first humanly-perceptible language and a second humanly-perceptible language is provided. As shown in FIG. 4, such a method may comprise providing a statement in a first humanly-perceptible language to an audience (step 200), and soliciting a translation of the statement from the audience in a second humanly-perceptible language and receiving the translation (step 202). In one embodiment, a statement may be made available, for example, via a website associated with a translation service hosting a translation model/engine as disclosed herein. In such instances, users of the website may, for example, indicate all languages in which they are fluent and can confidently provide translations, wherein such capabilities may be associated with a user profile. The users may access or otherwise be provided a statement(s), such as, for example, movie subtitles (which may be considered a proxy for conversational speech), television program scripts, music lyrics, or legal documents, for translation by such users from a first humanly-perceptible language to a second humanly-perceptible language, presuming that the users have at least some proficiency in the two languages. In this manner, a translation of the statement is solicited from various members of the audience (users), and any such translations provided by users are received by the website from the audience. In some instances, upon submission of the translation, a user may be presented with a subsequent statement for translation, wherein the user may potentially submit a limited or unlimited number of translations. A verification program may be implemented, in some instances, to verify the accuracy or level of sufficiency with respect to the translations provided by the audience, wherein only those translations confirmed as meeting a predetermined threshold criteria may be used to enhance/train/build the translation model/engine.

[0038] After submitting the translation, the statement in the first humanly- perceptible language (i.e., the statement to be translated) may be compared with a corresponding portion of the statement in the second humanly-perceptible language (i.e., the statement as-translated) so as to determine an inter-language or other correlation between the first and second humanly-perceptible languages, as represented by step 204. Such an inter-language correlation may be applicable to translate a further communication between the first humanly-perceptible language and the second humanly-perceptible language. Such embodiments of the present invention also serve to increase the quantity of translation data available to the translation model/engine, which, in turn, may serve to enhance the translation model/engine and improve the quality, accuracy, and efficiency of subsequent translations. Furthermore, this approach may allow for development of domain specific correlations or dictionaries, as used such machine translators, to provide high accuracy and sufficiency translations of communications specific to a particular domain such as, for example, law, medicine, and military.

[0039] In yet another aspect of the present invention, a method for improving translation of a communication between a first humanly-perceptible language and a second humanly-perceptible language is provided. As shown in FIG. 5, the method may comprise soliciting one of a translation of a first communication, in a first humanly-perceptible language, between the first humanly-perceptible language and a second humanly-perceptible language, and an evaluation of the first communication in the second humanly-perceptible language, from a translator and receiving the one of the translation and the evaluation, as represented by step 300. That is, in some instances, a member of an audience may submit or otherwise provide a statement to, for example, a website for translation by a human-translator between a first humanly-perceptible language and a second humanly-perceptible language. In other instances, the member of the audience may submit or otherwise provide a statement to the human-translator in the second humanly-perceptible language for evaluation of the accuracy of the correlation with the statement with respect to the first humanly-perceptible language. That is, the member of the audience may either solicit translation of the statement from a human-translator or solicit verification or other evaluation that a translation is accurate, wherein the human-translator may provide corrections or other comments as needed. [0040] Accordingly, the translated or evaluated statement in the second humanly- perceptible language may be compared to the statement in the first humanly-perceptible language for determination of an inter-language or other correlation between the first and second humanly-perceptible languages, as represented by step 302. Such a correlation may be used to enhance/build/train a translation model/engine such that the correlation may be applicable to translation of a second communication between the first humanly-perceptible language and the second humanly-perceptible language. That is, as more translation data is collected from the translated statements, the evaluated statements, and/or the evaluations, the translation model/engine incorporating such data can continually improve the quality and accuracy of translations provided thereby. [0041] In still yet another aspect of the present invention, a method for improving translation of a communication between a first humanly-perceptible language and a second humanly-perceptible language is provided. As shown in FIG. 6, the method may comprise collecting communication data from a first communication in a first humanly-perceptible language and a second humanly-perceptible language and associating the communication data with a general domain or context, as represented by step 400. For example, content of the first communication available in the first and second humanly-perceptible languages may be collected and stored in, for example, a first content repository associated with a general domain or context, wherein the content in the first content repository extends across various domains or contexts such as, for example, law, medicine, and military (i.e., the content is collected in connection with various scenarios).

[0042] The communication data may be evaluated with respect to a domain threshold criteria, as represented by step 402. For example, the communication data may be scored against a domain- or context-specific reference by using a computer algorithm. In such a manner, some or all of the communication data can be associated with one or more particular domains or contexts if respective threshold correlations can be determined. A portion of the above-threshold communication data in the first humanly-perceptible language may then be compared with a corresponding portion of the above-threshold communication data in the second humanly-perceptible language so as to determine a specific domain or contextual correlation between the first and second humanly-perceptible languages, as represent by step 404. In some instances, the above-threshold communication data in the first humanly-perceptible language and/or the above -threshold communication data in the second humanly-perceptible language may be compared with a corresponding specific domain reference so as to evaluate the first communication against the respective humanly-perceptible language. That is, the communication may be evaluated against a reference or standard to measure the accuracy of the inclusion of the communication in the specific domain or context. [0043] In an exemplary embodiment, the first communication associated with the first content repository may be compared against a predetermined threshold criteria for inclusion in a second content repository associated with a specific domain. That is, if the content of first communication is, for example, determined to be law-related, then it will be included in the translation data associated with the translation model/engine specific for translating a second communication determined to be associated with the domain of law, which may be accomplished using, for example, computer algorithms for scoring and/or weighting the translations. In contrast, if the content of the first communication is, for example, determined to be medical-related, then it would not be recognized by the translation model/engine specific as being associated with the domain of law, but would instead be included with the translation data associated with the translation model/engine as being specific to the domain of medicine. In this manner, domain specific information may be filtered from the general domain repository to create specific domain repositories used to enhance/build/train a translation model/engine for translations of communications determined to belong in a particular domain or context.

[0044] According to a further aspect of the present invention, a method for improving translation of a communication between a first humanly-perceptible language and a second humanly-perceptible language is provided, wherein the first and second humanly-perceptible languages are associated with a common character script. Such an approach may be particularly useful for translating a communication between a Latin character-based script, such as, for example, English, French, and Spanish, and a non-Latin character-based script such as, for example, Chinese, Japanese, and Greek. [0045] Previous translation schemes/programs have required that the communication to be translated be provided in a traditional script (i.e., traditional

Chinese, Japanese, or Greek script) or, alternatively, translated to the traditional script. This can be problematic in that inputting the traditional script into a computer device employing a translation scheme/program requires the use of a keyboard device utilizing the traditional script. However, some native speakers of a non-Latin character based foreign language may use a computer device utilizing a Latin character-based keyboard (i.e., a so-called "QWERTY" keyboard or equivalent, where English characters form the basis of the communication written in a non-English language). In such instances, native speakers of the traditional script may type in the Latin-based script, wherein the traditional script can be transcribed to the Latin character-based script, for example, in a phonetic to the traditional script in the Latin character-based script. However, as mentioned previously, prior translation schemes/programs may not recognize or comprehend the transcribed text (i.e., a non-English communication phonetically or otherwise represented by English character script) and, thus, may require, for example, the use of multiple or multi-function keyboard devices (i.e., an additional keyboard may be implemented utilizing the traditional script such that the user could submit text in the traditional script for translation).

[0046] As an example, a string of Japanese characters may represent a word/phrase that can be spelled out / transcribed phonetically into Latin-based characters, by using a QWERTY or equivalent keyboard device. Such a transcription would not necessarily be recognized by prior translation schemes/programs. Thus, a user seeking translation of the Japanese characters would need a keyboard device utilizing traditional Japanese characters/script. In this manner, it would be advantageous to provide a translation scheme/program capable of receiving the transcribed text for translation or, alternatively, translating to the transcribed text (rather than the traditional script). Accordingly, the first and second humanly-perceptible languages may be related through a common character script, such as, for example, the Latin character-based script, associated with both. [0047] In some embodiments of the present invention, as shown in FIG. 7, such a method may comprise receiving a text communication comprising the common character script arranged according to a first humanly-perceptible language, as represented by step 500. The common character script of the text communication may be correlated with communication data in the first humanly-perceptible language, as represented by step 502 (i.e., the script of the communication correlated with a particular humanly-perceptible language). In some instances, the communication data may be collected by transcribing content from a first character-based script to a second- character based script, and then translating the content provided in the second character-based script from the first humanly-perceptible language to the second humanly perceptible language. For example, content provided in a Latin character- based script, though still in Japanese language and representing Japanese script, may be transcribed to or otherwise correlated with traditional Japanese script. After the transcription, the content may then be considered as being provided in the Japanese character script, wherein the transcribed content may then be translated from a first humanly-perceptible language (i.e., Japanese) to a second humanly-perceptible language (i.e., English). As an example, the content may be translated from the Japanese language (provided in the non-traditional script) to, for example, the English language. Alternatively, the content may be first provided in the English language and then translated to the Japanese language, possibly in the non-traditional Latin character- based script, wherein the content may then be transcribed to traditional Japanese script, if desired.

[0048] The correlated communication data in the first humanly-perceptible language may be compared with corresponding communication data in a second humanly-perceptible language so as to determine a correlation between the common character script of the text communication arranged according to the first humanly- perceptible language and the second humanly-perceptible languages, as represented by step 504. That is, the transcribed text may be compared with the translated text to determine a correlation therebetween. For example, the non-traditional Japanese text (i.e., in the Latin character-based script) may be compared with the translated text provided in the English language to determine the correlation therebetween. The correlation may then be applicable to translate a portion of a second text communication comprising the common character script arranged according to the first humanly-perceptible language and the second humanly-perceptible language. [0049] According to yet another aspect of the present invention, a method for customizing translation of a communication between a first humanly-perceptible language and a second humanly-perceptible language is provided. As shown in FIG. 8, the method may comprise associating at least one of personal information and demographic information of each of a translator and an intended recipient of a translated communication with a corresponding translation profile, as represented by step 600. In such a manner, the translations provided by the translator or received by the intended recipient may have such personal and/or demographic information associated therewith. In some embodiments, the translator and/or intended recipient may be solicited for information (which may be accomplished, for example, via a website or a computer network) pertaining to personal and/or demographic information thereof, such as, for example, language preferences, geographic location, age, occupation/profession, etc., which can then be associated with a name, email address, or other identifier of the translator / intended recipient for associating the personal and/or demographic information therewith. In such a manner, user or translation profile is created. [0050] Such a method may further comprise translating a communication between a first humanly-perceptible language and a second humanly-perceptible language by the translator, at least partially based upon the translation profile of the translator corresponding to the translation profile of the intended recipient of the translated communication (i.e., at least some commonality between the profiles of the translator and the intended recipient), as represented by step 602. That is, the match between the translator and intended recipient profiles is utilized to customize or specifically tailor the translation to the preferences of the intended recipient (i.e., the translator exhibits common characteristics with the intended recipient). For example, a communication may, in some instances, be transmitted in a first humanly-perceptible language by email such that the intended recipient receives the communication in the second humanly- perceptible language, with the translated communication having additional adjustments/modifications thereto so as to account for the preferences associated with the translation profile or other particular characteristics of the recipient. Thus, the communication may be tailored to the particular preferences or context/domain/demographic of the intended recipient. Similarly, when a translator submits a translation, the translator profile associated with the translator may be utilized to associate the translation with particular profile fields or preferences or characteristics associated therewith. [0051] According to still another aspect of the present invention, a method for facilitating customized translation of a communication between a first humanly- perceptible language and a second humanly-perceptible language is provided. As shown in FIG. 9, the method may comprise associating at least one of personal information and demographic information of a translator with a translation profile, as represented by step 700. For example, a translator may be solicited for information (which may be accomplished, for example, via a website or computer network) pertaining to personal and/or demographic information thereof, such as, for example, language preferences, geographic location, age, occupation/profession, etc., which can then be associated with the translator to create the particular translator's profile. Such a method may further comprise associating the translation profile with a translation between a first humanly-perceptible language and a second humanly-perceptible language provided by the translator, as represented by step 702. That is, a translation provided by the translator may be associated with the translation profile so as to form a correlation therebetween (i.e., domain- or context-specific). [0052] Such a method may further comprise determining an inter-language correlation between the first humanly-perceptible language and the second humanly- perceptible language, at least partially based upon the translation profile of the translator, as represented by step 704. In such a manner, the translation provided by the translator may be appropriately associated with the characteristics of that translator's profile. Accordingly, the profile-correlated translation may then be used for building and/or enhancing a translation model/engine such that a personalized translation can be provided to persons seeking translations, where such persons would have similar profiles/preferences/characteristics as the translator. That is, the inter-language correlation may further be applicable to translation between the first humanly- perceptible language and the second humanly-perceptible language of a communication for an intended recipient of the translated communication, wherein the intended recipient has a translation profile corresponding to the translation profile of the translator (for example, where the translation profile of the intended recipient comprises personal information and/or demographic information of the intended recipient, which corresponds with the personal/demographic information associated with the translator).

[0053] In an exemplary embodiment, an individual may be in the legal profession and, as such, may have a translation profile exhibiting such information. Accordingly, a translation submitted by the individual is associated with the legal profession such that the translation may be used to build/enhance/train a translation model in the specific domain or context of law. In this manner, the translator-specific information and translation data associated with translations provided by the individual may be used for customizing future translations, as performed by the computer device, of communications determined to be law-related.

[0054] According to another aspect of the present invention, a method for facilitating customized translation of a communication between a first humanly- perceptible language and a second humanly-perceptible language is provided. As shown in FIG. 10, the method may comprise analyzing communication data provided by a user so as to identify a preferred communication style of the user, wherein the preferred communication style includes, for example, a native dialect of the user, as represented by step 800. Such an approach may account for an individual's unique style of communication (driven by surroundings, home environment, education, etc.) and vocabulary in which the individual is most comfortable communicating. As such, embodiments of the present invention may be provided so as to permit the individual to send and receive communications in the native dialect and/or preferred communication style that takes into account, for example, the individual's background, native dialect, area of a particular country, etc. [0055] In some instances, the individual may be solicited for information such that a native dialect and/or a preferred style of communication thereof may be identified. In other instances, previous communications, such as, for example, email correspondence, may be used to identify the same or similar information of that individual. In any instance, analysis of this information to determine, for example, the native dialect or preferred style of communication may be based on computational linguistics techniques, such as, for example, Hidden Markov Models and n-gram probabilities. In some instances, an individual's profile may be continually refined by intermittently requesting or otherwise determining additional information therefrom after the initial solicitation. [0056] Such a method may further comprise associating the preferred communication style of the user with a translation between a first humanly-perceptible language and a second humanly-perceptible language, as represented by step 802. By associating the preferred style of the individual with the translation, the translation may be tailored to the preferences thereof, and may be applicable to both inter-language and intra-language "translations"). For example, if it is determined that the user prefers a southern U.S. dialect, then communications received by the individual can be modified to adjust certain words/phrases to the user-preferred southern U.S. dialect in which the user is more comfortable communicating. Conversely, communications sent by the individual can be recognized as including a southern U.S. influence, and this factor can be taken into account when providing a translation of the communication for an intended recipient, whether preferring such a southern U.S. dialect or not. Such a method may further comprise determining a correlation between the first humanly- perceptible language and the second humanly-perceptible language, at least partially based upon the preferred communication style of the user, as represented by step 804. The correlation may be applicable to communications directed to and from the user involving translation between the first humanly-perceptible language and the second humanly-perceptible language.

[0057] According to still yet another aspect of the present invention, a method for facilitating text communication is provided. Such a method may be utilized to translate text communications between different styles, wherein the text communications may include, for example, emails, documents, attachments, and any other textual content. For example, an email communication written in an informal style may need to be converted into a professional style before reaching an intended recipient. In such instances, the converted style may be highly or entirely predicated on the preferences of the intended recipient. Furthermore, such an approach may be particularly useful for stylistic translations of real-time communications across agencies, wherein each agency has a different communication protocol and language specific thereto. For example, sharing of information across government agencies, such as, for example, between the Federal Bureau of Investigation and the Central Intelligence Agency, has becoming increasingly important in combating terrorism within the United States. Accordingly, embodiments of the present invention may be useful in stylistically translating communications between the agencies such that they may be capable of more effectively and efficiently responding to real-time information. In some instances, translation models may be built/trained to accomplish such stylistic translations, as similarly performed by translation models/engines configured to translate between and/or within languages.

[0058] As shown in FIG. 11, such a method may comprise analyzing an original text to determine a language style associated therewith, as represented by step 900. As mentioned previously, the style may include, but is not limited to, formal, informal, professional, or any other style of text, grammar, etc. In some instances, the style of the original text may be determined by an appropriate computer algorithm or other analytic software. Such a method may further comprise analyzing data associated with an intended recipient of the original text to determine a language style associated with the intended recipient, as represented by step 902. That is, a style may be associated with the intended recipient such that the original text is stylistically translated to the preference of the intended recipient. Such a preference may be indicated by either the sender or the intended recipient. In any instance, such preferential data may be associated with the intended recipient so as to ensure the original text is communicated in the appropriate manner.

[0059] Such a method may further comprise converting the original text from the language style associated therewith to the language style associated with the intended recipient prior to forwarding the converted original text thereto, as represented by step 904. In some instances, the original text may be translated from a first humanly- perceptible language to at least a second humanly-perceptible language, at least partially based on one of the language style associated with the original text and the preferred style of the intended recipient. That is, in some instances, the original text may be translated between different styles as well as different languages during transmission of the original text.

[0060] Many modifications and other embodiments of the inventions set forth herein will come to mind to one skilled in the art to which these inventions pertain having the benefit of the teachings presented in the foregoing descriptions and the associated drawings. For instance, certain embodiments are premised on the communication being available in a humanly perceptible language, non-inclusively such as Belarusian, Bulgarian, Czech, Chinese, Danish, Dutch, English, French, German, Italian, Japanese, Korean, Norwegian, Polish, Portuguese, Romanian, Russian, Serbo-Croatian, Slovak, Spanish, Swedish, and Ukrainian. However, in some instances, the source and destination languages used with the communication network may be languages that are not humanly perceptible, for example, where communication is between machines operating on different protocols. Of course, the systems and methods disclosed herein may also have and/or require associated computer program products for accomplishing the necessary functionality, and such is considered to be within the scope of the present invention, as will be appreciated by one skilled in the art. Therefore, it is to be understood that the inventions are not to be limited to the specific embodiments disclosed and that modifications and other embodiments are intended to be included within the scope of the appended claims. Although specific terms are employed herein, they are used in a generic and descriptive sense only and not for purposes of limitation.

Claims

THAT WHICH IS CLAIMED:
1. A method for improving translation of a communication between a first humanly-perceptible language and a second humanly-perceptible language, comprising: collecting communication data from a first communication in a first humanly- perceptible language and a second humanly-perceptible language; and comparing a portion of the first communication in the first humanly-perceptible language with a corresponding portion of the first communication in the second humanly-perceptible language so as to determine an inter- language correlation between the first and second humanly-perceptible languages, the inter-language correlation being applicable to translate a portion of a second communication between the first humanly- perceptible language and the second humanly-perceptible language.
2. A method according to Claim 1, wherein the first communication is a single literature work.
3. A method according to Claim 1, wherein the inter-language correlation is a statistical translation model of a statistical machine translator.
4. A method according to Claim 1 further comprising comparing at least one of the portion of the first communication in the first humanly-perceptible language and the portion of the first communication in the second humanly-perceptible language with a corresponding language reference so as to determine an accuracy of the first communication in the respective humanly-perceptible language.
5. A method according to Claim 1, wherein collecting communication data from a first communication further comprises: providing a statement in the first humanly-perceptible language to an audience; and soliciting a translation of the statement from the audience in the second humanly-perceptible language.
6. A method for improving translation of a communication between a first humanly-perceptible language and a second humanly-perceptible language, comprising: providing a statement in a first humanly-perceptible language to an audience; soliciting a translation of the statement from the audience in a second humanly- perceptible language and receiving the translation; and comparing a portion of the statement in the first humanly-perceptible language with a corresponding portion of the translation in the second humanly- perceptible language so as to determine a correlation between the first and second humanly-perceptible languages, the correlation being applicable to translate a portion of a communication between the first humanly-perceptible language and the second humanly-perceptible language.
7. A method according to Claim 6, wherein the correlation is a statistical translation model of a statistical machine translator.
8. A method according to Claim 6, wherein a plurality of translations of the statement in the second humanly-perceptible language is received from the audience and the method further comprises comparing the plurality of translations of the statement with a reference translation so as to evaluate each of the plurality of translations.
9. A method for improving translation of a communication between a first humanly-perceptible language and a second humanly-perceptible language, comprising: soliciting one of a translation of a first communication in a first humanly- perceptible language between the first humanly-perceptible language and a second humanly-perceptible language, and an evaluation of the first communication in the second humanly-perceptible language, from a translator and receiving the one of the translation and the evaluation; and comparing a portion of the first communication in the first humanly-perceptible language with one of a corresponding portion of the first communication translated into the second humanly-perceptible language by the translator and a corresponding portion of the first communication in the second humanly-perceptible language evaluated by the translator so as to determine an inter-language correlation between the first and second humanly-perceptible languages, the inter-language correlation being applicable to translate a portion of a second communication between the first humanly-perceptible language and the second humanly-perceptible language.
10. A method according to Claim 9, wherein the inter-language correlation is a statistical translation model of a statistical machine translator.
11. A method for improving translation of a communication between a first humanly-perceptible language and a second humanly-perceptible language, comprising: collecting communication data from a first communication in a first humanly- perceptible language and a second humanly-perceptible language and associating the communication data with a general domain; evaluating the communication data with respect to a domain threshold criteria and associating above-threshold communication data with a specific domain; and comparing a portion of the above-threshold communication data in the first humanly-perceptible language with a corresponding portion of the above-threshold communication data in the second humanly-perceptible language so as to determine a specific domain correlation between the first and second humanly-perceptible languages, the specific domain correlation being applicable to translate a portion of a second communication, associated with the specific domain, between the first humanly-perceptible language and the second humanly-perceptible language.
12. A method according to Claim 11, wherein the inter-language correlation is a statistical translation model of a statistical machine translator.
13. A method according to Claim 11 further comprising comparing at least one of the portion of the above-threshold communication data in the first humanly- perceptible language and the portion of the above -threshold communication data in the second humanly-perceptible language with a corresponding specific domain reference so as to evaluate the first communication against the respective humanly-perceptible language.
14. A method for improving translation of a communication between a first humanly-perceptible language and a second humanly-perceptible language, the first and second humanly-perceptible language being associated with a common character script, the method comprising: receiving a text communication comprising the common character script arranged according to a first humanly-perceptible language; correlating the common character script of the text communication with communication data in the first humanly-perceptible language; and comparing the correlated communication data in the first humanly-perceptible language with corresponding communication data in a second humanly- perceptible language so as to determine a correlation between the common character script of the text communication arranged according to the first humanly-perceptible language and the second humanly- perceptible language, the correlation being applicable to translate a portion of a second text communication comprising the common character script arranged according to the first humanly-perceptible language and the second humanly-perceptible language.
15. A method according to Claim 14, wherein the common character script is an English language QWERTY character script.
16. A method according to Claim 14, wherein the correlation is a statistical translation model of a statistical machine translator.
PCT/US2009/062030 2008-10-27 2009-10-26 Method for translation of a communication between languages, and associated system and computer program product WO2010062542A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
US10873808 true 2008-10-27 2008-10-27
US61/108,738 2008-10-27

Publications (1)

Publication Number Publication Date
WO2010062542A1 true true WO2010062542A1 (en) 2010-06-03

Family

ID=41572365

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2009/062030 WO2010062542A1 (en) 2008-10-27 2009-10-26 Method for translation of a communication between languages, and associated system and computer program product

Country Status (1)

Country Link
WO (1) WO2010062542A1 (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9864745B2 (en) 2011-07-29 2018-01-09 Reginald Dalce Universal language translator
US9916306B2 (en) 2012-10-19 2018-03-13 Sdl Inc. Statistical linguistic analysis of source content
US9954794B2 (en) 2001-01-18 2018-04-24 Sdl Inc. Globalization management system and method therefor
US9984054B2 (en) 2011-08-24 2018-05-29 Sdl Inc. Web interface including the review and manipulation of a web document and utilizing permission based control
US10061749B2 (en) 2011-01-29 2018-08-28 Sdl Netherlands B.V. Systems and methods for contextual vocabularies and customer segmentation

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050228643A1 (en) * 2004-03-23 2005-10-13 Munteanu Dragos S Discovery of parallel text portions in comparable collections of corpora and training using comparable texts

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050228643A1 (en) * 2004-03-23 2005-10-13 Munteanu Dragos S Discovery of parallel text portions in comparable collections of corpora and training using comparable texts

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
A. KUMARAN, K. SARAVANAN, M. SANDOR: "WikiBabel : Community Creation of Multilingual Data" WIKISYM '08, [Online] 10 September 2008 (2008-09-10), pages 1-10, XP002566345 Porto, Portugal Retrieved from the Internet: URL:http://wiki-translation.com/BabelWiki08?bl=y#Fully_reviewed_papers> [retrieved on 2010-01-29] *
KOEHN P: "Europarl: A Parallel Corpus for Statistical Machine Translation" INTERNET CITATION, [Online] XP002351175 Retrieved from the Internet: URL:http://www.iccs.informatics.ed.ac.uk/p koehn/publications/europarl-mts ummit05.pdf> [retrieved on 2005-10-25] *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9954794B2 (en) 2001-01-18 2018-04-24 Sdl Inc. Globalization management system and method therefor
US10061749B2 (en) 2011-01-29 2018-08-28 Sdl Netherlands B.V. Systems and methods for contextual vocabularies and customer segmentation
US9864745B2 (en) 2011-07-29 2018-01-09 Reginald Dalce Universal language translator
US9984054B2 (en) 2011-08-24 2018-05-29 Sdl Inc. Web interface including the review and manipulation of a web document and utilizing permission based control
US9916306B2 (en) 2012-10-19 2018-03-13 Sdl Inc. Statistical linguistic analysis of source content

Similar Documents

Publication Publication Date Title
US7383542B2 (en) Adaptive machine translation service
US7756708B2 (en) Automatic language model update
Schultz et al. Multilingual speech processing
US6996520B2 (en) Language translation system and method using specialized dictionaries
US20080133245A1 (en) Methods for speech-to-speech translation
US20050216253A1 (en) System and method for reverse transliteration using statistical alignment
US20100138402A1 (en) Method and system for improving utilization of human searchers
US20110099000A1 (en) Method and device for transliteration
US20010029455A1 (en) Method and apparatus for providing multilingual translation over a network
US20060206481A1 (en) Question answering system, data search method, and computer program
US20050137854A1 (en) Method and apparatus for evaluating machine translation quality
US20110307241A1 (en) Enhanced speech-to-speech translation system and methods
US20080077406A1 (en) Mobile Dictation Correction User Interface
US8468149B1 (en) Multi-lingual online community
US8145472B2 (en) Language translation using a hybrid network of human and machine translators
US20040044517A1 (en) Translation system
US20090055381A1 (en) Domain Dictionary Creation
US20050209844A1 (en) Systems and methods for translating chinese pinyin to chinese characters
US20060149551A1 (en) Mobile dictation correction user interface
US20080208567A1 (en) Web-based proofing and usage guidance
US20110137636A1 (en) Context aware back-transliteration and translation of names and common phrases using web resources
Waibel et al. Multilinguality in speech and spoken language systems
US20140303961A1 (en) Systems and Methods for Multi-User Multi-Lingual Communications
US20070118357A1 (en) Word recognition using ontologies
US20050149499A1 (en) Systems and methods for improving search quality

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 09747960

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase in:

Ref country code: DE

122 Ep: pct app. not ent. europ. phase

Ref document number: 09747960

Country of ref document: EP

Kind code of ref document: A1