DEVICE FOR ANALYZING NATURAL LANGUAGE INCREMENTALLY, ADAPTIVE ANSWERING MACHINE AND METHOD USING THE DEVICE

Various modification of the present invention refers to various embodiment thereby, the cold air flows that can apply which may have bar, specific drawing illustrated in the embodiment are defined in the description are disclosed and. rapidly and to reduce a memory. However, the present invention with a particular embodiment of the physical shape not defined to be, included within the scope of the present invention all changing a concept and techniques, including replacement water and equalization should understood. Each drawing while describes similar references in a similar was to use components.

Number 1, number 2, A, a set of terms, such as B describes various elements which may be used; however, said components are said terms is don't is defined by. Said terms are components of one an object from other components is carried out by using an acidulous only. For example, rights of the present invention without a wireless type through a wire number 1 number 2 component can be designated components, similarly number 2 number 1 component elements can be designated. And/or a substrate having a number of associated term of items combination or plurality of associated a substrate of items includes which item.

To other components is any component "is connected" know "is connected" when-mentioned that, that different structural elements thereof connected directly to or may be is connected, intermediate the other components may be present that. to be understood. While, to other components is any component "is directly connected with the" know "are directly connected" when-mentioned that, intermediate the other components there is no will should be understood.

The present application only a term use in a particular embodiment used to describe the thereby, the cold air flows are added, is not intending to be defining the present invention. Contextually representation a plurality of differently it is apparent that without the carelessly, includes multiple representations. In the present application, "comprising" or "having." a set of terms, such as a specification to the features, number, step, operation, components, component or a combination of these is designates the feature to which is present does, number to execute another aspect of one or more, step, operation, components, component or a combination of these existence of a without excluding the possibility or additionally pre should understood.

Other is not defined, technical or scientific for a term including to the all terms are person with skill in the art in the present invention is in the field of the upwardly urged by equivalent to those that would have been understood have the meanings of wet liquid to flow down. Generally are defined as the dictionary used for, such as terms are on wherein the nodes refining the context of related techniques consistent semantics and having having the meanings must be interpreted to, the present application, become manifest in a do not define, excessively or is ideal for the widest sense of the formal does not interpreted.

First, the present application used in. off at the first and the second briefly terms.

Natural language ( nature ) those natural language or a language and acts as an-pro, having the activity of providing an artificial tan to made a [...] and is general outline for calling a seperately. All human speech activity to based, which is characterized, at universal of wet liquid to flow down. Russian, Korean as by country or people including a language written to the various natural language..

Morpheme (morpheme) of a morpheme other binding specifically with when realized in the him (allomorph) bimorph. Most morpheme to name a few often a morpheme since the is embodied in the form of also defined a class of bimorph. Morpheme with a unique identification is being subjected to the phoneme the combination of the meanings defined viewing appropriate do not go. Morpheme that magnetron harness sometimes descriptor (syntax) represents the relationship of the widest sense of means for written which, in addition to close-photograph the grammar case an room to guide the voice recognizer is included therein,.

Just, orthography ( accidence ) on a morpheme in the present invention in addition to speech for morpheme analyzer of a morpheme is's introduced may comprise an. For example english grammarthe national script law or morpheme free on the even morpheme terminated if necessary for morpheme analyzer, dummy morpheme help an infant to easily make, use can be made of,. Morpheme-up table which next approximately the amount of for classification system exhibits and exemplary.

In the present invention using morpheme information column relationship between combinations of the signal of a represent syntactic personality, . information from the detected text information. Morpheme analysis syntax a relation information i.e. relationship between non-is information which can be found. Morpheme analysis on thus field of orthography ( accidence ) if activity belonging to, syntax analysis syntactic field of activity is belonging to (semantic) of a reflector according to each step will. Morpheme relationship between non-next-up table which approximately the amount of information classification system exhibits and exemplary.

Hereinafter, the present invention according to a preferred embodiment by issuing an thereby, the cold air flows rapidly and to reduce a memory reference to drawing..

Also Figure 7 shows a graded according to one embodiment of the present invention natural language dialogue device (200) and a construction details. is a block account for element. First, gradual natural language dialogue device (200) overall behavior of, which together significantly simplifies the process also for 7 to illustrate the reference to time as large as that of.

Gradual natural language dialogue device (200) the, natural language utterance ( speech ) (adaptively) adaptively to device in an interactive response, natural language utterance character sensing method for translating a generating units gradually and digital signal processor, and a gradual transition in voice recognizer (210) and a, voice recognizer (210) gradually character received from by decomposing unit morpheme units after character morpheme dummy behind the morpheme the end of unit gradually adds, morpheme of morpheme information and speech generating the relation information between a gradual transition of non-target language analysis unit (220) and a, speech associated information character based on the information the unit gradually, corresponding to unit gradually character a gradual transition in that determine the response the interactive manager (230) can be constructed including.

Character gradually unit, tool for measuring to completion of includes morpheme essential required whether no signal is applied, contains not less than at least one may be proposed.

Gradually interaction manager (230) the, (hierarchical) hierarchical relation information (sub tree) subtree of a morpheme represented as the number of dummy morpheme including for analyzing other sub-tree of the sub-trees for analyzing than can be after accomplishment ( [...] ).

The problems of existing dialog system for enabling response that gradually natural language analysis device, i.e. gradually of non-target language analysis unit (220) is used. User's utterance acids to epoxygenated fatty acids therein response gradually surge control unit assists the main control unit as well as the use of entire utterance, during user's utterance even is checked as to whether the inputted utterance, i.e. entire utterance using of the forefoot of the present invention the user's utterance by a rope. predictive understanding. Gradually Figure 7 of non-target language analysis unit (220) using natural language dialogue device (200) concerns a.. Gradually of non-target language analysis unit (220) for meanihgfui efficiency of gradual voice recognizer (210) is have required either. Gradually of non-target language analysis unit (220) a gradual transition in utilizing the results of interactive manager (230), or the like, that has gradually make the necessary, is designed not negative isolation wall. Interactive embrace a region that other device therefore gradually gradually when of non-target language analysis unit (220) can be nutritive advantage of the benefits of both high.

Gradual natural language dialogue device (200) is prepended continues has an utterance SMS (IU, Incremental Unit) unit gradually according addition of morpheme analysis unit (221, also 7 the not shown, shown in 2 also) and a syntax analysis unit (224, also 7 the not shown, also shown in 2) is updated is. Morpheme analysis unit (221), syntax analysis unit (224) the design, each when the units gradually characters or analysis newly all, on the another portion contains film-like or can be updated only portion.

Gradual natural language dialogue device (200) method for gradually voice recognizer (210) has utterance the end even before completed until a given until and a open converts. Such voice recognition is repeatedly executes short period previous voice recognition is executed the character voice later voice previous starting to write is delayed unit gradually the outcome of the recognition character ells and gradually can be in addition to then units.

Gradually of non-target language analysis unit (220) morpheme analysis the striker bar during voice recognizer (210) is shifted results firing control unit assists the main control unit and in view of the unacknowledged so morpheme respective appeared in one morpheme can be S100.

Gradually of non-target language analysis unit (220) of syntax analysis morpheme the results of the analysis control unit assists the main control unit maintaining ignition morpheme is stored in the virtual account database unacknowledged so can be and analyzing relationships between. Text message having sentence simultaneously been terminated, otherwise then coming out of a morpheme of speech can be for estimating what the.

Gradually of non-target language analysis unit (220) is based on information obtained until intermediate firing of the entire utterance, and may act as predictors of intended. User of silence even structure sentence is parsing control unit assists the main control unit utterance of a user via a user transceiver when the controller determines that was not fire silence end of sentence without the present until the very end of a sentence is structure a continue analysis.

Gradually of non-target language analysis unit (220) user's silencing using feedback to the sensing part measures a user for storing alternate recognition results to ink by making the touch soft language understanding module (120) level kinds of an accurate flow rate over the rest of user via the feedback can be leading the urine of utterance.

Gradually interaction manager (230) also gradually when therefore, the overall processing can be used to speed up or.. As the is also refers to 6 also to 2 at description below a reference to.

Syntax analysis unit (224) for training corpus is based on existing corpus for automatically made possible by the fact that the (also 7 the not shown, also shown in 2). Established, using a simple sentence the corpus is required that have not been input yet but information the morpheme. a locking feature may be added. Corpus is trained syntactical analysis unit (224) that have not been input yet the tool for measuring, the types of morpheme can be predicting termination.

Also Figure 2 shows a graded according to one embodiment of the present invention the new device language analysis is a block account for components.. Also Figure 3 shows a morpheme analysis unit (221), Figure 4 syntax analysis unit (224) of a fitting pin to explain the is exemplary. Figure 5 shows a gradual morpheme analyzer also, gradual Figure 6 for parsed corpus graded for the learning, the learning unit (225) of a fitting pin to explain the is exemplary. 6 also to 2 also gradual reference to the new device language analysis behavior of a component. off at the first and the second.

Gradual language analysis device a, natural language dialogue system (100) gradually in (incremental) in device analyzing language, unit gradually to make a character morpheme (morpheme) (IU, Incremental Unit) a split unit, dummy behind the morpheme the end of unit gradually character morpheme (dummy) the adding of morpheme number 1 analysis unit (222) and a, speech of morpheme and a dummy morpheme each decomposition of a (part of speech) for morpheme by determining a parts-of-speech for making morpheme number 2 analysis unit (223) and a, speech morpheme based on the information relationships between data syntax analysis unit (224) can be constructed including.

Gradual language analysis device a, stored in a memory by decomposing unit morpheme a corpus (corpus) corpus gradual morpheme corpus from parts-of-speech information and for a gradual transition in storing relation information corpus for the learning, the learning unit (225) further includes, morpheme corpus for parts-of-speech information associated based on the information, morpheme number 2 analysis unit (223) for of a morpheme has generate unfiltered reconstructed information and parts-of-speech, syntax analysis unit (224) has of a morpheme between relation information can be produced. Further gradual corpus for the learning, the learning unit (225) the, behind the morpheme the end of corpus gradual adds morpheme dummy corpus, stored in a memory included in corpus of a morpheme of morpheme dummy corpus corresponding to speech may determine are replaced by of.

Morpheme number 2 analysis unit (223) the, character gradually tool for measuring units morpheme essential required to completion of which has both a dummy if the program is judged to be the morpheme and determines the speech indicative of termination sentence, otherwise the masks the morpheme speech representing a morpheme essentially can be deciding.

Morpheme number 2 analysis unit (223) the, each decomposition of a morpheme of speech are replaced by of morpheme dummy based on the information can be determining.

Morpheme number 2 analysis unit (223) the, character morpheme based on unit gradually for parts-of-speech information method in modifying some or all of the, character a following-to the grafting unit gradually character morpheme unit gradually are replaced by determines, as a parting line, and dummy morpheme of, a following-morpheme for calibration information for the parts-of-speech can be produced. Further syntax analysis unit (224) the, morpheme based on calibration information speech relationships between information between morpheme elongated to a method can be produced relation information.

Syntax analysis unit (224) the, dummy morpheme of character based on speech to determine whether terminated sentence unit gradually or, morpheme relationships between information with a predetermined condition concerning a are met composition of unit gradually characters depending on whether the predetermined region of terminated. Further the predetermined condition, at least more than one of said morpheme each decomposition of a morpheme into a test mode the morpheme when the relation information text gradually and assumed to be it is a sentenceamerica it concluded units, each decomposition of a morpheme any one of the dummy the morpheme when which has no relation information the text gradually it is a sentence terminated units can be assumed to be. Yet further syntax analysis unit (224) the, character gradually sentence non-terminating units when a decision is made that a vehicle, character input unit gradually after predetermined time or more character gradually unit gradually character a following-to the grafting unit is not inputted a feedback (feedback) to the grafting unit gradually if the corresponding feedback determining syntax so as to output the output of voice signal syntax for controlling can be voice synthesis device.

Gradually of non-target language analysis unit (220) in morpheme analysis unit (221) the input to a PGA may not terminated text message having sentence. A sentence is terminated whether is not and non-terminating any it should morpheme next to be passed through to the morpheme is turned on and, with gradually horizontal pattern and the of non-target language analysis unit (220) are defined. of.

Attaching and fixing the timer to the morpheme analyzer (121) and a sentence structure analyzer (122) has an intact sentence structure, that is for analyzing. Thus is gradually the analysis utterance existing relationship with the derma when applied analyzer of a morpheme analysis of the firing is interrupted may be possible but yet the suspension arm is not firing the tool for measuring prediction and the morpheme prediction of whether terminated. is impossible. Thus gradual language analysis device of the present invention analyzer of training structure and from existing method, training corpus to be used should is improved of the.

Of a morpheme analyzer gradual

Morpheme analyzing unit of the present invention (221) (morpheme number 1 analysis unit (222) and a morpheme number 2 analysis unit (223) composed) user's utterance is inputted during gradual voice recognizer (210) as a result of character gradually divided into a predetermined number of files units until and a gradually inputted text with included in and uttered using units of a morpheme. has an electric compressor for compressing refrigerant. Input the striker bar during voice recognizer (210) gradually character as a result of unit, output the character string is divided into a unit each morpheme each noun classification, are replaced by of morpheme irradiation is information representing parts-of-speech.

Gradually of non-target language analysis unit (220) in morpheme analysis unit (221) text message having sentence terminated the input to a PGA which may not, case input utterance behind the morpheme is further and will be can be means. Thus morpheme analysis unit (221) generates a setting input before special a symbol, for example adding "" .. For example "her at number 7 now a drama" as input when receiving a telephone call via a morpheme analysis unit (221) the "her at number 7 now a drama" analyzing a sentence as. One "" can be considered a morpheme. I.e., an actual firing the "" but does not include dummy for morpheme analyzer is morpheme.

Morpheme analysis unit (221) each morpheme the trained method as. to be located at the upper part. Firing is interrupted result of a morpheme by information "" morpheme, i.e. after the morpheme of speech is ignited automatically be predicted. "" When-charge signal followed by a sentence and with an intact, i.e. "" morpheme for obtaining the firing are substantially end of sentence when the means of the kind comprising a be tagging speech ( english grammar or the national script law on the although we cannot same can be highly entangled terminated sentence refers to). Also "" end of this sentence was the input utterance is which a sentence is an intact, firing control unit assists the main control unit. occurred.

Morpheme Figure 3 number 2 analysis unit (223) is shown operation of. Gradual text message having sentence of a morpheme analysis ends without waiting for can be initiates analysis. Inputted text with preprocessor comprises a unit that the data is segregated into according to each morpheme preprocessor comprises a parts of speech are awarded tag.

Process interpreting gradually over time of the wireless terminal exists since are opened upward and downward entered ad having a wrong information parts-of-speech analyzing already displays the been. In this case before should can modify the information. Morpheme purified from number 2 analysis unit (223) the text gradually entire each when the units or analysis again updated according to an analysis only or part may be selected for method. When analyzing entire again is portion connects the vacuum a natural this problem. Analyzes only part and updating the application time to alter content before are replaced by again analyzing. is required additional functionality.

Morpheme analysis unit (221) input the striker bar during voice recognizer (210) amount measuring sensor the output of the output of the morpheme analysis unit (221) is necessary that there is a change in the structure of. Gradual voice recognizer (210) has word unit, or time unit output and disconnecting the unit gradually character may be selected a structure. Morpheme analysis unit (221) the certain reference and digital signal processor, and gradually unit is made according to a method corresponding to find out method can be designed. Gradual voice recognizer (210) unit gradually character unit time speech where the output has enable signal when enabled as words similar to the word received from intact and cannot ensure. For example when short-range in units of character gradually gradual voice recognizer (210) the " effect a" of a character "as present" unit into gradually. Thus morpheme analysis unit (221) stores previously incoming speech information of a morpheme of a morpheme should it can be modified.

The gradual parsed

Syntax analyzing unit of the present invention (224) has of a morpheme analysis unit (221) of a morpheme is analyzed using part-of-speech information by of morpheme relationship between found relation information can be produced. Syntax analysis unit (224) of the existing sentence structure analyzer (121) the information is reproduced in the point and the other comes into gradual since data structure may be modified continues even before is completed sentence structure of sentence as to form a mass to be copyright 2000 whether a message.

Syntax analysis unit (224) connected to an existing sentence structure analyzer (121) morpheme unlike necessarily used to indicate a relationship between of a morpheme is still may not been input. For example, "I see fun movies" sentence as "interesting" object for in is "movie". However gradual natural language dialogue device (200) in the search unit on an "I was yoked to a pleasing" in these incoming when syntax analysis unit (224) the "interesting" morpheme set in the quartz carrier part can't find yet.

In this regard syntax analysis unit (224) has after the coming out of a morpheme, the types of the grammar should, for example, to predict a serves. "Interesting" in on subsequent morpheme nouns may be, "interesting" of information to be processed will be a corresponding to the prediction is.

To this end a morpheme analysis unit (221) added in one can analyze "" together. "" Through automatic analysis of "" and of a morpheme input until now, and may act as predictors of relationship. Also sub having for a "" "" rely on the that tree be estimated.

Gradually interaction manager (230) also gradually when therefore for analyzing tree sub been done and transmitted down running on the first "" for analyzing tree, a subsystem contained with the is the relegating the non first performed processing for processing to be removed over the entire surface can be used to speed up or.. Dummy morpheme a "" is, a subsystem contained with the associated morpheme dummy tree presence of morpheme Image generator generating an Image including a semantic is, required to completion of tool for measuring a top plate of a morpheme less since been input is means.

Figure 4 syntax analysis unit (224) operation of the planar sheet to exhibit a process, syntax analysis unit (224) has of a morpheme analysis unit (221) "" added in each of a morpheme includes morpheme relationship between may be described.

Syntax analysis unit (224) also morpheme analysis unit (221) similarly to the new sound recognition result is through voice recognition result previous when receiving a telephone call via a syntactical analysis result error of. may be discovered. In this case even gradual morpheme analyzer (121) method such as a solve the problem on computing system. Morpheme analysis unit (221) alternatively from syntax analysis unit (224) each morpheme tag relationship morpheme and morpheme as well as must also modified.

Syntax analysis unit (224) of the sentence structure, and may act as predictors of. Thus current utterance which the it does not or sentence is interrupted can be for estimating. Text message having sentence terminated opposite direction of the spring force direction grammatically portion syntax analysis unit (224) of silence user behind silence even followed by firing the others would not finished by the wireless terminal assumes that the analysis may be.

Sentence termination is ignited subjected to a predicting a morpheme not may utilize "". Morpheme analysis unit (221) in installed on "" outside of the tool for measuring tag when the tag which means that, for example when a sentence is terminated silo terminated sentence drops when they break, the water can be present. Other method at "" for a bobbin having necessarily is that yet the text message having sentence un-is ignited not been terminated can be present.

A silencing not effect a termination of the utterance is analyzed part (224) the feedback symbols for the user can determine the. "You," feedback " [...] ?" to be generic and is versatile, such as it and short "where to yaw?", "what to same shall?" such as remaining as feedback to kinds of accurate leading utterance of. is capable of both. Length of silence user the kind of feedback, sentence structure would following the type of morpheme expected can be according to.

For example, "TV viewing controls mass information today I horse field" the present invention to an utterance as natural language dialogue device (200) is a degree of short "you" a feedback of the fixing pin for storing alternate recognition results to ink the people the red a subjects are linked to. formed by growing a naturalness. In addition "today at two o'clock movies--" that, an utterance is silence long enough for to behind the verb he will come, to predict "what to same shall?" as feedback to a user when further what kind of information from the detected text information the previously written, i.e. including a verb in example on fact that say can be recognition again. Such feedback device language analysis stopped electronically for a sentence in the display 18. only the analysis of the structure.

Gradual of a morpheme analysis and parsed for training corpus for

For existing analyzer entire sentence corpus is formed inside the tie. an analyte using a. Corpus is trained utilizing the is ends when utterance morpheme analysis of the utterance may be possible but yet the morpheme of the suspension arm is not end of and articles of which prediction methods produces prediction speech conversion of whether.. Thus gradually of non-target language analysis unit (220) to be used for training corpus. favor I need to ask of different from existing corpus. However new corpus is made are discharged to current traffics information and then receives a response from the same line on which necessary based corpus according to the value of the films and automatic corpus is.

Gradual corpus of a morpheme analyzer for existing corpus and to incrementally one sentence included in voice recognizer (210) to make users identical as a consequence of "" behind the sentence in each can be of are chosen from zeolite, plastic. Figure 5 gradual corpus using of a morpheme analyzer example for making.. "" Morpheme the kind original type of morpheme ol subsequently corpus are in accordance.

Gradual parsed corpus for access same also can be produced according to method. Gradual Figure 6 new parsed corpus. example for making. Morpheme if any lower part of the inner panel is "" morpheme the incoming yet and none when a virtual root node can be linked to "0".. Tool for measuring after the "" also. can be linked to "0".

Also according to one embodiment of the present invention Figure 8 shows a natural language dialogue graded detailed and method is flow account for step. Also 8 natural language dialogue gradual reference to detailed and method. off at the first and the second step.

Natural language dialogue gradual the method, an adaptive responsive natural language for ignition an interactive method in process the device, to make a character gradually split unit morpheme units (S820), morpheme the end of unit gradually character morpheme dummy behind the morpheme for adding an analysis steps number 1 (S830) and a, each decomposition of a morpheme and a dummy morpheme determines, as a parting line, are replaced by of morpheme for number 2 morpheme for making parts-of-speech analysis steps (S850) and a, speech morpheme based on the information syntax for making relationships between (S860) analytical can be constructed including.

Natural language dialogue gradual the method, prior to said number 1 morpheme analysis steps (S830), natural language utterance character sensing method for translating a gradually and digital signal processor, and a gradual transition in generating units further includes a voice recognition step (S810) can be constructed.

Natural language dialogue gradual the method, after (S860) analytical syntax, associated information speech character based on the information (S870) the unit gradually, corresponding to unit gradually character a gradual transition in that determine the response the step managing dialogue includes (S880), gradual managing dialogue step to a layer-represents (S880) has the relation information of a morpheme subtree dummy morpheme including for analyzing other sub-tree of the sub-trees for analyzing than may after it will accomplish.

Natural language dialogue gradual the method, stored in a memory unit morpheme corpus is gradual by decomposing morpheme corpus from corpus parts-of-speech information and for a gradual transition in storing relation information (S840) step is further includes a learning corpus, morpheme corpus for parts-of-speech information associated based on the information, morpheme number 2 for preprocessor comprises a analytical generate unfiltered reconstructed information and parts-of-speech, parsed between morpheme step can be produced relation information.

Morpheme number 2 (S850) analytical the, character gradually tool for measuring units, including both morpheme essential required to completion of when the dummy the morpheme the sentence indicative of termination and determines the speech, otherwise required to the morpheme masks speech representing a morpheme can be deciding.

Morpheme number 2 (S850) analytical the, character morpheme based on unit gradually for parts-of-speech information method in modifying some or all of the, character a following-to the grafting unit gradually character morpheme unit gradually are replaced by determines, as a parting line, and dummy morpheme of, a following-morpheme for calibration information for the parts-of-speech can be produced.

Syntax analytical the (S860), dummy morpheme of character based on speech to determine whether terminated sentence unit gradually or, morpheme relationships between information with a predetermined condition concerning a are met composition of unit gradually characters depending on whether the predetermined region of terminated.

Gradual natural language dialogue method as the graded reticle natural language dialogue progressively device language analysis device (200) and overlaps a a description does not substrate longer since.

Although in some sides are described in terms of device is perfectly suitable for, description of such sides are corresponding method that indicates which clear, wherein device. corresponding to method steps. According to request particular implementations, are of the invention in the embodiment can be implemented in hardware or software. Program code are of the present invention in the embodiment, method one for the implementation of a program which operates in a computer program having code can be performed as a product.

Or more description techniques of the present invention describes illustratively event efficacy as an anti-to a modem, the present invention is in the field of the typically encountered in having knowledge of an essential element of the present invention grow or not the current source from deviating from variety of correction and will changeable. Thus the present invention to a disclosure slimly examples embodiment of the present invention is supported by the upper case and and to limit the for described, in the embodiment of the present invention by such slimly range of limited not. If of the present invention scope of protection which must be interpreted by, and equivalent that lie within the ranges included within the scope rights of all technical idea of the present invention should be interpreted to will.