Automatic or semiautomatic translation system and method with post-editing for the correction of errors

• US 20030040900A1
• Filed: 07/26/2002
• Published: 02/27/2003
• Est. Priority Date: 12/28/2000
• Status: Active Grant

• Current Assignee: Dʼagostini, Giovanni
• Original Assignee: Dʼagostini, Giovanni
• Inventors: DʼAgostini, Giovanni
• Granted Patent: US 7,580,828 B2
• Time in Patent Office: Days
• US Class Current: 704/2
• CPC Class Codes: G06F 40/47 Machine-assisted translation, e.g. using translation memory

Abstract

A computer translation system having viewing means on monitor of a translation-interface involving a couple of text-columns, wherein each text-column includes parallel scrolling field means, characterized in that said couple of text-columns is divisible horizontally, to realize at least two couples of parallel scrolling fields, forming at least temporarily a “+”-structure in which, one pair of scrolling fields is below for translation and correction after translation, forming a couple of main translation and correction fields (Fl, B-U) and the second pair is over these for accumulation of what translated checked and corrected, forming a couple of accumulating fields (A1,U/Z) and wherein, said translation system has transfer means (Memline) to progressively transfer the couples of top paragraphs of said main translation and correction fields (F1, Z) to said accumulating fields (A1, Z).

Claims

1. A computer translation system having viewing means on monitor of a translation-interface involving a couple of text-columns, wherein each textcolumn includes parallel scrolling field means, characterized in that said couple of text-columns is divisible horizontally, to realize at least two couples of parallel scrolling fields, forming at least temporarily a “

+”

-structure in which, one pair of scrolling fields is below for translation and correction after translation, forming a couple of main translation and correction fields (F1, B-U) and the second pair is over these for accumulation of what translated checked and corrected, forming a couple of accumulating fields (A1,U/Z) and wherein, said translation system has transfer means (Memline) to progressively transfer the couples of top paragraphs of said main translation and correction fields (F1) to said accumulating fields (A1).

2. A computer translation system according to claim 1, having viewing means on monitor of a translation-interface involving a “

T”

structure, where the upper bar of the “

T”

is a line of control buttons and the vertical shank is a dividing line forming a couple of text-columns, wherein each column includes scrolling field means proportioned in the width to the length of the text contained in rapport to the opposed parallel one, characterized in that said couple of text-columns is divided horizontally forming at least - structure for having four scrolling fields in a form of at least two pairs of parallel scrolling fields, one pair below for translation and correction after translation, forming a main couple of translation and correction fields (F1, B-U) and over this a second pair for accumulation of what translated checked and corrected, forming a couple of accumulating fields (A1,U), and wherein evidencing means of modified text in said couple of main translation and correction fields is further provided.

18. A method according to any of the previous claims, characterized by providing at least a checking window of the autolearning with three checking lines, correction and input (E):

one as fragment of phrase of the text to translate;

one as fragment of phrase translated in a correct way, object of the correction in progress or already carried out;

one as programming semantic code.

19. A method according to any of the previous claims, characterized by providing a system supplying the in-line consulting of dictionary with alternative translations (5).

20. A method according to any of the previous claims, characterized by providing during the correction in post-editing:

determining the position of the cursor in the correction area or otherwise if memorized a portion by evidencing, calculating automatically the number of the paragraphs and corresponding sentences and words of the translated document, from the origin and,

on the base of the correspondence of the dotting positions, supply in a window in the screen with;

the corresponding portion of the sentence of the document to translate,

the sentence portion previously evidenced in the correction area concerned individuated by the presence of the cursor from the last correction or previously evidenced (H), and

in order to allow to the operator to delimit evidencing the fragment of phrase corresponding to that interested by the correction (I).

23. A method according to any of the previous claims, characterized by providing a translation system comprising upwards and downwards parallel scrolling means for both fields in the same time of each couple.

24. A method according to any of the previous claims, characterized by providing:

a transformation control means from capital to small letters,

a transfer control means between fields and to a Data File,

a copy control means,

a paste control means,

a word counter means,

a lock control means,

a spelling control means,

a “

cut returns”

control means,

a “

START-translation”

control means including, automatic or semiautomatic option;

a paragraph by paragraph (“

LineByLine”

) machine-aided translation control means,

an interactive-option translation control means (“

TQ”

), where each fragment between punctuation is submitted to the operator for check and correction, before recording;

a check and correction control means to allow after an automatic translation;

the transfer in said couple of accumulating fields (A1,U) the checked and corrected sentences from said couple of main translation and correction fields (F1, B-U), and

the retranslation of at least the remaining first paragraph in said couple of main translation and correction fields (F1, B-U).

3. A computer translation system according to previous claims characterized for having paragraph-counter means to show the number of remaining paragraphs to check during checking and correction.

4. A computer translation system according to previous claims, characterized by having means to show for an additional couple of scrolling fields below said couple of main translation and correction fields (F1, B-U) to include in one or the other the text portion to translate (L1,U).

5. A computer translation system according to previous claims, characterized by having correction-autolearning means and retranslation means to retranslate at least the first paragraph to check in said couple of main translation and correction fields (F1, B-U).

6. An automatic/semiautomatic translation method, according to previous claims, for translating text from one language to another, of the type utilizing a system which comprises means of automatic or semiautomatic translation and means to allow the correction of the translation operated by the computer and means to implement sentence/word dictionary/ies into the computer system, and word process means in post-editing for check and correction the mistakes of what has been translated with eventual activation of automatic search and replace, characterized by:

integrate the said post-editing in a single ensemble with the translation system, and

identify and memorize the correction that is carried out in post-editing, and

activate the search of said mistakes and replace with said corrections in the said translated text, and

activate the memorizing of said correction in said translation system with integration of one or more program semantic codes,

in order to obtain in the same moment, directly connected to the respective correction, not only the extension of same correction to the text in post-editing, but also, in the same time, the autolearning of the correction itself to the said translation system.

7. An automatic/semiautomatic translation method, for the translation in the desired language, according to previous claim, of the type utilizing a system that includes:

at least one electronic dictionary for correct translations of words, sentence-fragments and sentences, associated for each standard translation to a program semantic code, that activates during the translation of a text or text portion (F1, B-U), a respective adapting program of the translation in progress to the preceding text portion and/or to the following text portion in function of the specific features;

means for making the translation, phrase by phrase, of the text to translate with sequential analysis of the following words and/or sentence-fragments within the extent of treatment of the respective sentence, associated to comparing means of said words or sentence-fragments, to individuate materially or virtually also with declination or base or root, their presence into said electronic dictionary (1,

     2), to make possible a sequential translation;

processing means of the translation on the base of said dictionary and of the corresponding said program semantic codes;

with eventual adapting of the current translation to the preceding text portion and to the text portion that follows that in translation-processing and,

with eventual adapting in conformity of the current translation also of the part already translated;

output means to memorize and show the translation made;

characterized by providing furthermore;

a plurality of said program semantic codes (soms, sofs, somsn, sofp, verbinfcon, verbinfcon2, ag4 ax4), also variously combined together as a language, each code or each portion code being associated part to a respective translation software with different modality of translation;

said modality being known to the operator that operates the translation in order to allow the use of one or other code or the combination of more codes to define a certain translation software for a specific text portion or sentence fragment;

memory means of the carried out corrections (F1, B-U) and, on request of the operator;

activation means of an autolearning program as far as immediately previously corrected (E),

determining the new teaching to operate comprising;

the text portion that contains said correction, and

the corresponding original text portion that had to be translated, and

imputing one or the combination of more than one of said program semantic codes to this new teaching for adding in said electronic dictionary (1, 2, 3, 4,

     5), in order to be able to be reutilized immediately in the following translation.

8. A method according to previous claims, characterized by having:

activation means, on command, after the correction, of the following steps;

individuate the last change made on the text comparing the original previous text and the modified text;

if indicated (E), propose in a dialog window the previous text and the modified text or the text to be modified (D);

on acceptance (OK), make an automatic search and replace in the whole text already translated.

9. A method according to previous claims, characterized by providing at least the option of the fully automatic or semiautomatic translation (A).

10. A method according to previous claims, characterized by providing proportioning means for automatically proportioning the width in said text-columns couple to the length of the contained text, after each translation or retranslation.

11. A method according to previous claims, characterized by having a system that provides different writing, about missing words after translation.

12. A method according to previous claims, characterized by providing automatic charging and unloading means for the text to translate and for the translated text respectively, fragmenting them, in couples of memory fields (C1.C2.C3.C4.C5) horizontally placed in head on the respective control-bar.

13. A method according to previous claims, characterized by providing that said accumulating fields (A1,U) is displaceable from above the said couple of main translation and correction fields (F1, B-U), to said control-bar and vice versa.

14. A method according to previous claims, characterized by providing for each correction carried out in the translated text, besides the integrated search replace options and of autolearning, the autoactivation of the differentiation of the characters of the word, corrected fragment and corresponding portion of the text to translate, putting them in different characters (F).

15. A method according to previous claims, characterized by providing a system that includes the possibility to impute a different teaching (5) depending on context (electric, electronic, chemical, mechanic, aeronautic, medicine, sporting, clothing, legal, spoken), each able to be activated by a domain code previously assigned before translation.

16. A method according to previous claims, providing translation, paragraph by paragraph, from a bottom fields couple to said main couple, and after correction and autolearning (LineByLine, A-M) passing the corrected text to said couple of accumulating fields (A1,U).

17. A method according to previous claims, characterized by providing means which make possible after automatic translation phase, the sequential post-editing with correction and checking by a skilled person, in autolearning and search-replace mode, transferring progressively the couple of the first checked paragraphs from said couple of main translation and correction fields (F1, Z), to said couple of accumulating fields (A1, Z) and automatically retranslating the remaining to check, using previous teachings.

25. A Method according to claim 17, characterized in that after transfer the checked top paragraphs text-couple from the said main couple of fields (F1-Z) to the bottom of said accumulating-fields couple (A1), said automatic retranslation is operated on the first top paragraph of the said main couple of fields (F1-Z) of the remaining text to check, in order that in the respective retranslation, waiting time, the operator is allowed to read the previous translation and prepare oneself for amendments after issue of said retranslation in the first paragraph.

26. A Method according to claim 25, characterized in that before checking paragraph by paragraph, said retranslation is made only if the interested paragraph to check contains text portions already subjected to previous amendments and autolearning, made between the full automatic first translation and the moment to check.

21. A method according to previous claims, characterized by providing a translation system in which the said electronic dictionary is double, one for one direction and one for the opposite direction (S) and in which both have a series of fields:

for searching the single words to meet during translation (1),

for the following text portion the word under consideration (2);

for the base standard translation (3);

for the program semantic code-s (4) and

for alternative translations personalized with operative sectorial codes (5).

22. A method according to previous claims, characterized by providing a translation system in which the control bar of the main interface, has at least one visualization control for making appear one or more couples of said fields or vice versa.