Data processing: speech signal processing – linguistics – language – Linguistics – Translation machine
Reexamination Certificate
1999-09-23
2001-10-09
Thomas, Joseph (Department: 2644)
Data processing: speech signal processing, linguistics, language
Linguistics
Translation machine
C704S004000
Reexamination Certificate
active
06301554
ABSTRACT:
BACKGROUND OF THE INVENTION
Since the time when improvements in transportation began to significantly reduce the inconvenience and cost of cross-border travel, the desirability of universal communication has been recognized. In the 1960s, for example, international efforts were made to promote Esperanto as a universal language. While that effort ultimately failed, the large number of fluent speakers—between 1 and 15 million worldwide—and the scope of the efforts illustrate the problem's importance. Esperanto did not succeed because it required acquisition of both a new grammar and a new vocabulary, the latter posing a far greater challenge for would-be speakers.
The improving ease and speed with which information can now be transmitted worldwide has augmented the need for universal communication. Current efforts have focused most heavily on automated translation among languages. Systems now in use generally store, in a source and a target language, millions of frequently used words, phrases and combinations, relying for accuracy and robustness on the occurrences in the text to be translated. Such systems are by definition incomplete, since no system can possibly store every possible word combination, and their usefulness varies with the linguistic idiosyncracies of their designers and users. It is almost always necessary for a human to check and modify the output translation. These systems also translate one word at a time (and so operate slowly), and require a separate database unique to each target language. Moreover, because they are programmed to recognize distinctive language characteristics and their unique mappings from one language to another, each translation must be done individually. In other words, the time required for multiple translations is the sum of the times for each translation performed individually.
Translation is difficult for numerous reasons, including the lack of one-to-one word correspondences among languages, the existence in every language of homonyms, and the fact that natural grammars are idiosyncratic; they do not conform to an exact set of rules that would facilitate direct, word-to-word substitution. It is toward a computational “understanding” of these idiosyncracies that many artificial-intelligence research efforts have been directed, and their limited success testifies to the complexity of the problem.
U.S. Pat. No. 5,884,247 (issued Mar. 16, 1999) describes an approach toward language translation in which natural-language sentences are represented in accordance with a constrained grammar and vocabulary structured to permit direct substitution of linguistic units in one language for corresponding linguistic units in another language. The vocabulary may be represented in a series of physically or logically distinct databases, each containing entries representing a form class as defined in the grammar. Translation involves direct lookup between the entries of a reference sentence and the corresponding entries in one or more target languages.
In accordance with the '247 patent, sentences may be composed of “linguistic units,” each of which may be one or a few words, from the allowed form classes. The list of all allowed entries in all classes represents the global lexicon, and to construct an allowed sentence, entries from the form classes are combined according to fixed expansion rules.
Sentences in accordance with the '247 patent are constructed from terms in the lexicon according to four expansion rules. In essence, the expansion rules serve as generic blueprints according to which allowed sentences may be assembled from the building blocks of the lexicon. These few rules are capable of generating a limitless number of sentence structures. This is advantageous in that the more sentence structures that are allowed, the more precise will be the meaning that can be conveyed within the constrained grammar. On the other hand, this approach renders computationally difficult the task of checking user entries in real time for conformance to the constrained grammar.
SUMMARY OF THE INVENTION
In accordance with the invention, the constrained grammar is defined in terms of allowed sentence types (rather than in terms of expansion rules capable of generating a virtually limitless number of sentence types). In this way, it is possible to easily check user input (word by word, or in the form of an entire document) for conformance to the grammar, and to suggest alternatives to sentences that do not conform.
As in the '247 patent, the present invention provides an artificial grammar for expressing the thoughts and information ordinarily conveyed in a natural grammar, but in a structured format amenable to automated translation. The allowed sentence types are sufficiently diverse to permit expression of sophisticated concepts, but, because sentences are derived from an organized vocabulary according to fixed rules, they can be readily translated from one language to another. Preferably, the vocabulary is represented in a series of physically or logically distinct databases, each containing entries representing a form class as defined in the grammar. Translation involves direct lookup between the entries of a reference sentence and the corresponding entries in one or more target languages.
Accordingly, to use the invention, a natural-language sentence is either drafted in accordance with the allowed sentence structures, or translated or decomposed into the (typically) simpler grammar of the invention but preserving the original vocabulary. User input may be received via an editor programmed to follow the user's inputted text for conformity with the allowed sentence structures.
The translated output is as easily understood by a native speaker of the target language as the input was to the author of the original text. Thus, it is possible to carry out “conversations” in the grammar of the invention by formulating statements according to the grammar, passing these to an interlocutor for translation and response, and translating the interlocutor's responses. For example, a business person native to the United States and without knowledge of German can conduct a meeting with native German speakers using as a translation device a laptop computer configured in accordance with the invention, exchanging thoughts via the computer. Indeed, the same thoughts may be simultaneously broadcast to multiple interlocutors each speaking a different language, with their individual responses simultaneously and multiply translated as well. Correspondents can exchange messages by e-mail, in their native languages, simply by formulating the messages in accordance with the invention's grammar; recipients speaking different languages who have e-mail systems implementing the invention receive the message translated into their native languages, and their responses are automatically translated into the original sender's language upon their arrival; in this way, each correspondent is exposed only his or her native language.
The invention is advantageously employed even in situations demanding an ultimate output in a natural language, since translation to this format is readily accomplished. For example, a news reporter might file a story worded in the invention's grammar for dissemination to numerous bureaus serving different national audiences. The story is instantly translated into the appropriate languages upon arrival at the different bureaus, where it may then be further refined into a form suitable for communication to the audience. The skills required if further translation is desired are essentially editorial in nature, and thus require less specialized training than would be necessary, for example, for true language translation; indeed, communications media already employ personnel to carry out the similar tasks of editing and revising raw news material taken from wire services.
REFERENCES:
patent: 4829423 (1989-05-01), Tennant et al.
patent: 5020021 (1991-05-01), Kaji et al.
patent: 5237502 (1993-08-01), White et al.
patent: 537167
Testa Hurwitz & Thibeau L
Thomas Joseph
WordStream, Inc.
LandOfFree
Language translation using a constrained grammar in the form... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Language translation using a constrained grammar in the form..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Language translation using a constrained grammar in the form... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2614918