Fujitsu Corporation

Atr Interpreting Telecommunications Research Labs

Akira Ushioda

Seiji Okura

Masaru Fuji

Kenichi Utsumi

Nagaaki Koshino

Miyozo Maeda

Ken-ichi Itoh

Yasuyuki Goto

Itaru Shibata

Kozo Sueishi

Tatsuo Yamashita

Tomoki Nagase

Koji Tsukamoto

A processor performs semantic analysis on a query and generates one or more semantic structures where each structure is expressed by a graph. The processor generates retrieval keys corresponding to combinations of nodes connected directly or indirectly in the semantic structures, in addition to retrieval keys corresponding to minimum units of semantic connections between nodes in the generated semantic structures. The processor retrieves relevant documents whose sentences are matched to combinations of nodes, by using the generated retrieval keys, in the semantic structures stored in an index for retrieval on a database storing the documents.

Retrieval device and method

A typical sentence having a specific typical characteristic in the sentence is divided. A division target typical sentence is divided on the basis of a small clause definition. The sentence is divided where positions suitable for dividing the typical sentence based on the structure are expressed by a user. A small clause string including small clauses that serve as independent sentences is created after the division. The small clause string is compared to the structure patterns, and a structure pattern that is determined to match the small clause string is selected as a result of the typical sentence division.

Computer-readable medium having sentence dividing program stored thereon, sentence dividing apparatus, and sentence dividing method

A unified-evaluated-value computing unit quantitatively combines a translation probability between words computed by a corpus-based statistical-value computing unit, dictionary information computed by a dictionary-information referring unit, and syntactic structure information computed by a syntactic-structure-value computing unit, to compute a unified evaluated-value as an index for phrase alignment.

Computer product for phrase alignment and translation, phrase alignment device, and phrase alignment method

In a computer for supporting translation work, an acquisition unit retrieves a pair of original document including a plurality of original sentences and translated document including a plurality of translated sentences from a document storage unit. An estimation unit counts the number of original sentences included in the original document and the number of translated sentences included in the translated document, and estimates a position of a translated sentence corresponding to each original sentence in the translated document based on a ratio of the number of original sentences to the number of translated sentences and a positional order of each original sentence in the original document. An alignment unit aligns each original sentence with the translated sentence located at the estimated position, and outputs information indicating the correspondences between the original and translated sentences.

Translation supporting program, apparatus, and method

A translation support device for supporting a process of translating an original sentence in a language into another language, comprising a partial expression setting unit, a dummy head setting unit, an original sentence replacement unit, a translated sentence obtaining unit, and a translated sentence replacement unit.

Translation support device

A confirmation link edition unit receives a confirmation link specified by a user. A paragraph correlation unit respectively divides an English text and a Japanese text into a plurality of paragraphs according to the specified confirmation link. A segment correlation calculation unit correlates an English segment to a Japanese segment for each paragraph. A correlation edition unit provides a user the correspondence obtained by the segment correlation calculation unit, and edits the correspondence according to a correction instruction from the user if any.

Translation correlation device

An interface unit issues input/output instructions regarding an input of a translation target sentence, an output of a translated sentence, and a translation control. A machine translating apparatus translates a document in a certain language into a document in another language. A translation memory device translates a sentence by searching an original/translation database in which sentences in a certain language and original/translation sentences in another language corresponding thereto have been accumulated. A data compatible processing unit makes original/translation information translated by the machine translating apparatus and original/translation information translated by the translation memory device common, thereby enabling them to be fetched mutually as original/translation information.

Apparatus, method, and computer-readable medium for language translation

A control section is used which executes the following steps. That is, a) making a plurality of voters respectively vote evaluation of voting-object information in accordance with an on-line voting system by using a plurality of clients, each client being set at a voters sides, b) calculating average evaluation in accordance with voting results, c) deciding a voter's rank of each of the voters in accordance with a deviation between average evaluation and evaluation by the voter concerned, and d) determining overall evaluation of voting-object information at least in accordance with a result of weighting a voting result correspondingly to the voter's rank.

Evaluation apparatus with voting system, evaluation method with voting system, and a computer product

When a keyword described in Japanese is input, the input keyword is converted from Japanese into English. Thus, a retrieval keyword described in English is generated based on the input keyword described in Japanese. Documents that are described in English and that contain the retrieval keyword described in English are retrieved from a database. The retrieved documents are translated from English into Japanese. The documents translated into Japanese are compared with the input keyword described in Japanese. Thus, the validity of the results retrieved from the database is evaluated. Thus, even if the language of a database from which the data is retrieved is different from the language of the input keyword, retrieved results exactly based on the input keyword can be output.

Apparatus and method for retrieving data from a document database

An information searching apparatus converts an input key-term into key-terms for searching the database by at least two times of expansions using a related-term dictionary and a translation dictionary, and stores the history of the expansions in the form of a tree structure. Then, a part or the whole of the thus-stored tree structure is displayed on a screen. A searcher designates appropriate terms among the thus-displayed information as key-terms for searching.

Information searching apparatus for displaying an expansion history and its method

A token is attached to a word class sequence whose probability of appearance in text data is equal to or more than a predetermined value. A set of words and tokens included in a word/token sequence concerning the text data, is separated so that a probability of generation of the word/token sequence concerning the text data becomes the highest. The token is then replaced with a phrase included in the text data.

Word/phrase classification processing method and apparatus

In a word clustering apparatus for clustering words, a plurality of words is clustered to obtain a total tree diagram of a word dictionary representing a word clustering result, where the total tree diagram includes tree diagrams of an upper layer, a middle layer and a lower layer. In a speech recognition apparatus, a microphone converts an input utterance speech composed of a plurality of words into a speech signal, and a feature extractor extracts predetermined acoustic feature parameters from the converted speech signal. Then, a speech recognition controller executes a speech recognition process on the extracted acoustic feature parameters with reference to a predetermined Hidden Markov Model and the obtained total tree diagram of the word dictionary, and outputs a result of the speech recognition.

Class-based word clustering for speech recognition using a three-level

Recording and erasing optical information can be done by using an alloy film capable of forming two stable crystalline states differing in crystal texture and optical characteristics by being irradiated with optical energies under different conditions. A thin memory film preferably including 60 to 90 atom % of Indium (IN) and 10 to 40 atom % of Bismuth (Bi) is formed on a substrate.

Optical information memory medium including indium (In) and bismuth (Bi)

Recording and erasing optical information can be done by using an alloy film capable of forming two stable crystalline states differing in crystal texture and optical characteristics by being irradiated with optical energies under different conditions. The thin memory film preferably includes not more than 60 atom % of Gallium (Ga) and not less than 40 atom % of Antimony (Sb).

Optical information memory medium recording and erasing information

Recording and erasing optical information can be done by using an alloy film capable of forming two stable crystalline states differing in crystal texture and optical characteristics by being irradiated with optical energies under different conditions. The memory film includes 35-45 atom % of indium (In) and 55-65 atom % of antimony (Sb). The memory film may include an additional element M selected from the group consisting of Al, Si, P, S, Zn, Ga, Ge, As, Se, Ag, Cd, Sn, Te, Tl, Pb, Bi, and a combination of these elements, the composition of the memory film being expressed by the formula (In.sub.x Sb.sub.1-x).sub.1-y M.sub.y.

Growing community of inventors

Kawasaki, Japan

Akira Ushioda