DE102007003862A1 - National languages communication method for use between e.g. communicator and communicant, involves storing corrections made during sense of revaluation, input of planed text and output of received text in buffer - Google Patents

Abstract

The method involves understanding a text composed in a national language of a communicator (1.02) by a communicant (2.12) for understanding the text of a message, where the message is composed in a plan language (1.03) that is selected as a transfer medium for adaptation of a sense of meanings. The text is introduced for pre-translation and retranslation. Terms in technical texts are included into the text for translation processes from a technical dictionary. Corrections made during sense of revaluation, input of planed text and output of received text are stored in a buffer.

Description

Field of the invention

The The invention relates to a method for technical communication between Including people with different national languages a planned language. The process of a technical language translation is realized in such a way that the subjectively planned and understood meanings of meaning a messaging adapted to each other and from largely shielded from unwanted external influences become.

It It is differentiated that human communication is an individual subjectively planned, generated, transmitted, mediated, transmitted, received and subjectively understood process of messaging is. To adapt the understanding of meaning between subjective scheduled and generated messages and for adaptation received and subjectively understood messages make sense adjustments with supports technically guided and realized Dialogues both on the side of the communicator (author of the message) and the communicant (recipient of the message).

initial situation

According to Cölfen, the efficiency of machine translation systems is generally overestimated or misjudged. The raw translations created with machine translation systems must be followed by post-translations and text edits ( Cölfen, Hermann: Speech-based computer applications. In Applied Linguistics. Hg. Karlfried Knapp et al., Tübingen 2004: Francke. P. 172 ). Standardized communication, specialist dictionaries and so-called translation memory systems (TMS) can also facilitate specialist communications (cf. Göpferich, Susanne: Standardization of communication. In Applied Linguistics. Hg. Karlfried Knapp et al., Tübingen 2004: Francke, p. 457 ff ).

Proof that scientific texts can also be formulated by means of a planned language; is provided (see Pinter, Ana-Maria (ed.): Cybernetic Pedagogy. Klerigkibernetiko, Volume 11. Prague 1999: AL / Kava-Pech or Esterhill, Frank: Interlingua Institute: A History. New York 2000: Interlingua Institute ). The aim of a new automatic translation method is to improve the performance of previously known systems.

There Language is an individually generated means of communication can the performance of a translation system only be detected with indicators by those involved Be judged by users themselves.

In a technical translation system these are
linguistic expressiveness of planned messages,
linguistic unambiguity, as well as linguistic ambiguity should be avoided
simple dialogue during the generation of a message,
reliable transmission of the message and simple dialogue in the reception and understanding of a message.

The successful consideration of these performance indicators can be justified theoretically as a substitute. It is given a measure of improvement by, to what extent new translation system in combination with the necessary input and output dialogs designed so that the meaning of the translation process individual understandings of meaning can be adapted and an optimization of the technical translation system can take place.

A Clarification of the task then consists of the Communicator message to be transmitted in a technical guided dialogue, by word, by sentence or by sentence Correction notes adapt to linguistic ambiguities avoided, at least restricted. At the received Message are in reverse of this technically-led Dialogs to adjust the correction notes so that the subjective Understanding of meaning of the communicant is achieved. These Correction on the side of the communicant can be used as a post-translation be understood. But it can only be successful ambiguities in the transmitted message were avoided. Otherwise there is uncertainty in every correction process about what original meaning was conveyed with one Word, phrase or sentence was meant. Since this mutual adjustment and correction in the manner described so far is not sufficiently realized, the performance becomes Machine translation systems have generally been overestimated or incorrectly assessed.

Description of the state of technology

I. Corresponding patents / patent applications:

• EP 0 012 777 A1
• EP 0 244 871 B1 ( DE 37 51 716 T2 );
• EP 0 371 944 A2 ;
• EP 0 590 332 B1 ;
• EP 0 525 470 A3 ;
• EP 0 633 525 B1 ;
• EP 0 814 457 B1 ;
• EP 1 058 896 B1 ;
• EP 1 450 267 A3 ,

II. Disadvantages of the prior art

outgoing the meaningfulness of an intended message, a message transmitted and understanding a message can not be expected be that with a translation of a natural one Language in another a meaning adaptation of these different ones Meaning can be achieved. Even with a time-synchronous translation by a human translator increases the diversity of meaning contents between a communicator, as the source of the message and one or more communicants, as recipient of the message due to the risk of misunderstanding the translator. While a human translator for understanding the message still background information These are the possibilities offered by an automatic translator unavailable. To a limited extent, therefore, even in machine translations specialized dictionaries included. The ambiguities of individual terms and formulations, both in everyday language as well as in scientific languages, but can not be excluded. Furthermore Ambiguities are rarely transferable from one language to another. For machine-translated news, it is common to translate the "untranslated text", one related to the entire message, as possible self-contained sense meaning, is sought. The result is still unsatisfactory.

Next Individual and cultural differences remain to the individual in the sense understanding of a message. The avoidance of language cultural sense components can be solved even less by a machine translation are possible as to a limited extent in human translations is.

As as justification for the use of planned languages like Esperanto, Interlingua or Ido for the general Communication, provides a largely culture-neutral planning language the Possibility of meaning a message from a neutral cultural perspective. Prerequisite is a juxtaposition of divergent meaning meanings in a national language compared to those in a planned language. The previous automatic translation systems offer such juxtapositions not on.

Description of the invention

The invention is based on the object of a person subjectively expressed in a national language meaning meanings of a message of a second person or more other people in their respective national languages through technically guided interactive dialogues in the preparation of communication
in machine translation into a planned language through continuous back translation,
when sending the message and understanding the message
to facilitate.

The structure of the translation system ( 1 ) provides the option of being on the side of the newsreader or communicator ( 1:02 ) the entry of a planned message ( 1:01 ) using a speech coding from a national language N into a planning language P ( 1:03 ) in such a way that, by means of continuous back translation, an adjustment ( 1:05 ) by the national language culture ( 1:06 ) characterized meaning, which can then be translated as 1:04 ) and can be forwarded to the communicant.

On the side of the message taker or communicant ( 2.12 ) provides the structure of the translation system ( 1 ) the ability to understand a received message ( 2.14 ) using a speech coding from a planned language P into another national language N ( 2.13 ) so that an adaptation ( 2.15 ) by another national language culture ( 2.16 ) meaningful meanings can take place in such a way that they can then be understood as an understood message ( 2.11 ).

An overview of the conception is in 2 shown. Communication between a communicator and a communicant takes place through technically implemented speech coding ( 2.2 ) and speech decoding ( 2.3 ), which interculturally (with the respective national understandings of meaning) 2.5 and 2.6 ) can be adjusted.

The translation of the planned message in a planned language is done in several dialogue cycles word, sentence and sentence wise under control of the communicator in a method as defined in claim 1 and 2. The planned language used fulfills to a large extent the requirement of the linguistic expressiveness of planned messages, that is, it must contain a variety of expressions and be extensible, so that the meaning meanings of the national language of the communicator can be coded. Furthermore, the planned language used linguistically as clear as possible to be coded. That is, it should provide a coding scheme (word meanings, grammar, and sentence structures) in which the continuous back translation In this way, unique associations between words, sentences and sentences can be generated in such a way as to recognize, resolve and largely avoid linguistic ambiguities and national differences and ambiguities in a technically conducted dialogue. National-language ambiguities, as far as they are desired, are lost or must be resolved in the speech coding into several uniqueness.

conceptional unavoidable double and misinterpretations due to the diversity of linguistic Expressions can be restricted thereby by making corrections from both the communicator and the communicant saved by word and sentence and with repeated use be used as a basic data supplement. When appropriate Supplemented selection of the basic data is thus automatic Corrections possible.

An example of a functional sequence from the position of the communicator is shown as a flow chart in FIG 3 shown. in a normalized start state S ( 3.1 ), the system is ready to enter a text ( 3.2 ), for example, via the keyboard, and this in a buffer N ( 3.3 ). The system recognizes the national language N when entering the first words and confirms a changeable selection ( 3.4 ) the translation options between a national language N such as German, English, French, Spanish or Italian and a planned language P such as Esperanto, Ido or Interlingua. This selects a basic data set for encoding NP ( 3.5 ), which corresponds to a dictionary and the grammar rules. The entered parts of the text are translated into the planned language, word by sentence and sentence by sentence ( 3.6 ), issued ( 3.7 ) and in a buffer P ( 3.9 ).

The user as a communicator has at this point the choice of whether he wants to correct his input text or whether he wants to do without corrections. The same query also takes place after the text N and a back translation PN ( 3.10 ) has been compared in terms of words, sentences and sentences and the user has made a correction ( 3.11 ) of his text N in order to achieve a meaning adaptation. It is also possible to mark certain text areas that should not be translated as protected terms, phrases, names or citations. Also, by modification ( 3.12 ) the basic data ( 3.5 ) are supplemented with specialized dictionaries. The repetition of the same corrections can be restricted in the case of corresponding basic data settings by virtue of the fact that they are stored in terms of words and parts and thus automatic corrections are possible. While entering the correction and modification are in 5 screen displays shown in multitasking mode available. The level of view of the planned language translation is optimally limited to areas of application in which direct planned language communication is to be prepared and supported.

If corrections are omitted or if the corrections have been completed, output of the text via a printer or forwarding of the text in the plan request can be achieved by data transfer ( 3.13 ) are made according to claim 5. The end of the functional sequence forms a state E ( 3.14 ), which is usually taken after the routine queries have been made for the final storage of the text files recorded in the buffers and the selection and basic data settings.

An example of a functional sequence from the position of the communicant is shown as a flowchart in FIG 4 shown. In a normalized start state S ( 4.1 ), the system is ready to enter a text ( 4.2 ), for example as a transmitted file, and transfer this into a buffer P ( 4.3 ). The system recognizes in the analysis of the first words in which language P the text is written and confirms a changeable selection ( 4.4 ) the translation options between a planned language P such as Esperanto, Ido or Interlingua and a national language N such as German, English, French, Spanish or Italian. This selects a basic data set PN ( 4.5 ) containing a dictionary and the grammar rules for decoding PN. The adopted text is then translated into the national language, word by sentence and sentence by sentence ( 4.6 ), issued ( 4.7 ) and into a buffer N ( 4.9 ).

At this point, the user has the option of viewing the translation into his national language as complete or corrections ( 4.10 ). These corrections correspond to an adaptation of the text to his subjective sense or Fachverstennis. In addition to the text of the message in his national language, the text in the planned language is correspondingly or in 6 By way of example (English is chosen here as the national language), multitasking screen views are available. If, on the part of the communicant, the message has also been transmitted in his national language, the originally foreign national language can optionally be presented for comparison as an additional view and control level for comparison. For correction, the basic data ( 4.5 ) by modifying selection ( 4.4 .) are supplemented by specialist dictionaries. The corrections can be limited in the case of corresponding basic data settings, even in the case of meaning adjustments, by the fact that they are stored on the word and part-related basis when they occur for the first time, and automatic corrections are possible when they are repeated.

If corrections are omitted or the corrections have been completed, output of the text via a printer or a forwarding data transmission ( 4.12 ) are made according to claim 5. The end of the functional sequence is the state E ( 4.13 ) after the routine queries have usually been made for the final storage of the text files recorded in the buffers and the selection and basic data settings. The languages mentioned in the examples can be supplemented.

Advantages of the invention

With the method described in the invention is tried, the Performance of machine translation systems improve by a communicator (text writer) his intention when entering text, wordwise, sentence-by-sentence and sentence-by-sentence and change his text so that the in a planned language to the communicant (receiver) transmitted Text expresses this intention. Here we are a linguistic Transfer medium, a planned language, used as largely cultural neutral can apply, double meanings and linguistic idioms can recognized by the communicator when entering, corrected and thus largely be avoided. Further, the performance becomes improved by the communicant of a message this He receives text in a planned language and he with support of the translation system, the text to his subjective professional and national understanding of meaning in one wordwise, sentence-by-paragraph and sentence-by-phrase dialogue can (claim 1). The effort in this process can be appropriate individually required accuracies or understandings of meaning be adjusted.

The For this technically characterizable prerequisites are in the Description of the invention and in the claims described. It takes into account that the supplementary Possibilities of universal jobs, of advanced personal or multimedia computers as a single workstation or integrated into a local or universal transmission network at input, customization dialogs and output, also audio and video support not be excluded (claims 2 to 4).

QUOTES INCLUDE IN THE DESCRIPTION

This list The documents listed by the applicant have been automated generated and is solely for better information recorded by the reader. The list is not part of the German Patent or utility model application. The DPMA takes over no liability for any errors or omissions.

Cited patent literature

• EP 0012777 A1 [0008]
• EP 0244871 B1 [0008]
• - DE 3751716 T2 [0008]
• - EP 0371944 A2 [0008]
• - EP 0590332 B1 [0008]
• - EP 0525470 A3 [0008]
• - EP 0633525 B1 [0008]
• - EP 0814457 B1 [0008]
• - EP 1058896 B1 [0008]
• - EP 1450267 A3 [0008]

Cited non-patent literature

• - Cölfen, Hermann: language-related computer applications. In Applied Linguistics. Hg. Karlfried Knapp et al., Tübingen 2004: Francke. P. 172 [0003]
• - Göpferich, Susanne: Standardization of communication. In Applied Linguistics. Hg. Karlfried Knapp et al., Tübingen 2004: Francke, p. 457 ff [0003]
• - Pinter, Ana-Maria (ed.): Cybernetic Pedagogy. Klerigkibernetiko, Volume 11. Prague 1999: AL / Kava-Pech or Esterhill, Frank: Interlingua Institute: A History. New York 2000: Interlingua Institute [0004]

Claims (4)

Method for technically based communication between users with different national languages (communicators and communicants), characterized in that the users involved as communication participants have intelligent man-machine interfaces that have a dialogable information input and output, applications for the necessary information processing and optional may have an interface to a messaging system that the necessary information processing is supported by an operating system that allows multi-tasking the processing algorithms - multitasking - with parameter transfers, so that over it marks, cut and paste words, phrases and sentences the meaning adaptation and dialogue design the individual Requirements of users (communicators and communicants) can be adapted to that as codable and decodable transfer medium for the Nachr a planned language and optionally a national language is used. in that a planned language, for example Esperanto, Interlingua, Ido or another artificial language, is used as the codable and decodable transfer medium for the messages, that the required language can be supplemented with the planned language and ambiguities avoided by the communicator when entering a text Both by word, sentence, as well as sentence by sentence, constant forward and back translations between a national language and a planned language possible ambiguities avoided and the translated meaning can be adapted to the intended meaning and understanding that the communicant in the understanding of a text both by word, sentence, as well as sentence-wise, continuous forward and back translation between a planned and a national language possible double ambiguities avoided and so the translated meaning of meaning in the national language subjective sense and understanding adapted can be that the communicant for supplementary understanding the text of the message both in the selected as transfer medium planning language and the text written in the national language of the communicator for the extended adaptation of understandings and can include in the constant forward and back translations that selected words and text passages, names, terminology and citations can not be included in the translation, and terms in specialized texts for translations of supplementary specialized dictionaries can be inserted in the texts, that the changes made in the meaning of both the entry of the text and the delivery of the received text Corrections are saved and other similar or similar corrections can be made automatically. Method according to claim 1, characterized that Point-to-point messaging can take place where individual participants have access to the transmitted Have messages and participate in the communication. that Distribution messages can take place in which several participants have access to the transmitted Have news, that conference message submissions can take place where several participants or groups of participants Have access to the transmitted messages and yourself participate in the communication. Method according to claims 1 and 2, characterized that the forward and back translation and the input and output the news decentralized, under the control of the respective communication participants can take place that the pros and back translations including supplementary information processing distributed in a messaging system, that the forward and back translations including the complementary information processing anywhere throughout the communication system. Method according to 1, 2, and 3, characterized that input and output and transmission of messages both in printed form, in coded form by file, acoustically, or / and visually.