Showing papers on "Phrase published in 1994"

Semantic co-occurrence filtering for speech recognition and signal transcription applications

Abstract: A system and method for automatically transcribing an input question from a form convenient for user input into a form suitable for use by a computer. The question is a sequence of words represented in a form convenient for the user, such as a spoken utterance or a handwritten phrase. The question is transduced into a signal that is converted into a sequence of symbols. A set of hypotheses is generated from the sequence of symbols. The hypotheses are sequences of words represented in a form suitable for use by the computer, such as text. One or more information retrieval queries are constructed and executed to retrieve documents from a corpus (database). Retrieved documents are analyzed to produce an evaluation of the hypotheses of the set and to select one or more preferred hypotheses from the set. The preferred hypotheses are output to a display, speech synthesizer, or applications program. Additionally, retrieved documents relevant to the preferred hypotheses can be selected and output.

Lexical and message-level sentence context effects on fixation times in reading.

Abstract: Readers' eye movements were recorded as they read an unambiguous noun in a sentence context. In Experiment 1, fixation durations on a target noun were shorter when it was embedded in context containing a subject noun and a verb that were weakly related to the target than when either content word was replaced with a more neutral word. These results were not affected by changes in the syntactic relations between the content words. In Experiment 2, the semantic relations between the message-level representation of the sentence and the target word were altered whereas the lexical content was held constant. Fixation time on the target word was shorter when the context was semantically related to the target word than when it was unrelated. Intralexical priming effects between the subject noun and the verb were also observed. The results suggest that both lexical and message-level representations can influence the access of an individual lexical item in a sentence context.

Topic discriminator using posterior probability or confidence scores

Abstract: An improved topic discriminator includes an integrated speech recognizer or word and phrase spotter as part of a speech event detector, and a topic classifier trained on topic-dependent event frequencies. The event frequencies are determined from either or both transcribed data with a text event detector and untranscribed speech data with a speech event detector. In accordance with another aspect of the present invention, the phrase spotter is used to detect the presence of phrases without the need of parsing the output of a speech recognizer's hypothesized transcription. In accordance with another aspect of the invention, an improved technique is used to select a subset of the potential speech events on whose event frequencies the topic discrimination decision is made. Finally, in accordance with another aspect of the invention, an improved method of topic modeling is used to improve the performance of the topic discriminator.

Going in and out of Languages: An Example of Bilingual Flexibility:

Abstract: When bilinguals speak to one another, they choose a base language to interact in and then, depending on the need, code-switch to the other (guest) language for a word, a phrase, or a sentence During the perception of a code switch, there is a momentary dominance of base-language units at the onset of the switch, but it is unknown whether this base-language effect is also present in production, that is, whether the phonetics of the base language carry over into the guest language In this study, French-English bilinguals retold stories and read sentences monolingually in English and in French and bilingually in French with English code switches Both the stories and the sentences contained critical words that began with unvoiced stop consonants, whose voice onset times (VOT) were measured The results showed that the base language had no impact on the production of code switches The shift from one language to the other was total and immediate This manifestation of cross-linguistic flexibility is accounted for...

Phrase-final syllable lengthening and intonation in early child speech

Abstract: This research describes the development of phrase-final prosodic patterns in nine English-speaking children. The intonation feature of interest is the fall in the fundamental frequency of the voice...

A hierarchical stochastic model for automatic prediction of prosodic boundary location

Abstract: Prosodic phrase structure provides important information for the understanding and naturalness of synthetic speech, and a good model of prosodic phrases has applications in both speech synthesis and speech understanding. This work describes a statistical model of an embedded hierarchy of prosodic phrase structure, motivated by results in linguistic theory. Each level of the hierarchy is modeled as a sequence of subunits at the next level, with the lowest level of the hierarchy representing factors such as syntactic branching and prosodic constituent length using a binary tree classification. A maximum likelihood solution for parameter estimation is presented, allowing automatic training of different speaking styles. For predicting prosodic phrase breaks from text, a dynamic programming algorithm is given for finding the maximum probability prosodic parse. Experimental results on a corpus of radio news demonstrate a high rate of success for predicting major and minor phrase boundaries from text without syntactic information (81% correct prediction with 4% false prediction).

Speech recognition system

Abstract: PURPOSE:To provide the speech recognition system which performs high- precision recognition by using a language model which can express syntax and meaning relations wide in range than N-gram of words and can easily be structured from a text data base. CONSTITUTION:A speech vocalized while paused in phrase units is inputted to a phrase candidate output part 1, which outputs phrase lattices as the result of recognition in the phrase units to a sentence candidate selection part 2. A language model parameter storage part 3 is stored with a chain of two groups of adjuncts and the value of appearance probability corresponding to it as parameters of the language model. The sentence candidate selection part 2 selects an optimum phrase candidate sequence on the basis of the phrase lattices outputted from the sentence candidate output part 1 in consideration of the appearance probability of the chain of the two groups of adjuncts stored in the language model parameter storage part 3, and outputs the phrase candidate sequence as a sentence recognition result.

Parametric properties of numeral phrases in Slavic

Abstract: Numeral phrases in Russian display many unusual morphosyntactic properties, e.g., (i) the numeral sometimes assigns genitive (GEN-Q) to the following noun and sometimes agrees with it and (ii) the numeral phrase sometimes induces subject-verb agreement and sometimes does not. In this paper existing analyses of these properties are parametrized to accommodate related phenomena in other Slavic languages. First, Babby's (1987) proposal that GEN-Q is structural in Russian is shown not to extend to Serbo-Croatian, where it must be analyzed as inherent. Second, Pesetsky's (1982) idea that Russian numeral phrases may be either QPs or NPs also does not extend to Serbo-Croatian, where these are only NPs. This set of assumptions explains a range of seemingly unrelated facts about the behavior of numeral phrases in the two languages. Pesetsky's analysis is recast in terms of more recent hypotheses about phrase structure: (i) NPs are actually embedded in DPs and (ii) subjects are D-Structure VP-specifiers. Proposal (i) allows for a more explanatory analysis of GEN-Q assignment and proposal (ii) accounts for several distinctions between QP and NP subjects within Russian, also motivating the absence of these distinctions in Serbo-Croatian. Finally, it is shown that Polish can be assimilated to the proposed system.

Using cross-modal lexical decision tasks to investigate sentence processing.

Abstract: Recent investigations of sentence processing have used the cross-modal lexical decision task to show that the antecedent of a phonologically empty noun phrase (specifically, WH-trace) is reactivated at the trace position. G. McKoon, R. Ratcliff, and G. Ward (1994) claimed that (a) a design feature concerning the choice of related and unrelated targets is a possible confound in this work and (b) the conclusions drawn from this previous research are therefore called into question. These claims are considered in light of both McKoon et al.'s experimental findings and results of our own experiments in which we test their linguistic materials. We argue that their results may be due to the nature of their materials. Additionally, we argue that a follow-up experiment reported by G. McKoon and R. Ratcliff (1994) used a technique that is not comparable to the cross-modal lexical decision task. It is concluded that current evidence supports the claim that structural information is using during on-line sentence processing and that the cross-modal technique is sensitive to this.

The prosody of Greek speech

Abstract: Abbreviations 1. The Physiology of Prosody 2. The Syllable 3. Rhythm 4. Pitch 5. Word Prosody 6. Connected Speech 7. The Appositive Group 8. The Minor Phrase 9. The Major Phrase and Utterance 10. Topic and Focus Bibliography Index

The extraction of phrase structure during reading: Evidence from letter detection errors.

Abstract: In light of recent suggestions regarding the prominence of structure in speech production and comprehension, it has been postulated that structural processing might also play a similarly important role in reading Some evidence in support of this contention can be gleaned from eye-movement research However, more systematic support comes from recent work on letter detection during reading, which has shown that the rate of omission errors is inordinately high for morphemes that disclose phrase structure The results of three lines of research suggest that, early in text processing, readers attempt to extract a structural frame for the sentence to help the on-line integration of accessed representations, and that structure-supporting units recede to the background as the meaning of the sentence evolves

Tree-structured speaker clustering for fast speaker adaptation

Abstract: The paper proposes a tree-structured speaker clustering algorithm and discusses its application to fast speaker adaptation. By tracing the clustering tree from top to bottom, adaptation is performed step-by-step from global to local individuality of speech. This adaptation method employs successive branch selection in the speaker clustering tree rather than parameter training and hence achieves fast adaptation using only a small amount of training data. This speaker adaptation method was applied to a hidden Markov network (HMnet) and evaluated in Japanese phoneme and phrase recognition experiments, in which it significantly outperformed speaker-independent recognition methods. In the phrase recognition experiments, the method reduced the error rate by 26.6% using three phrase utterances (approximately 2.7 seconds). >

The prosodic basis of the tiberian hebrew system of accents

Abstract: 1 argue that the Tiberian system of accents which annotate the text of the Hebrew Bible has a prosodic basis. Accentual representations are constructed in terms of units somewhat similar to the modern prosodic hierarchy, and they deviate from syntactic constituency in ways that are characteristic of prosodic representations. They are con- strained by the effects of syntactic edges, geometric properties of prosodic phrases, principles for organizing phrases into higher-level constituents, and position in the phrase, associated with variations in tempo. It is shown that the Tiberian representation can best be understood by integrating results of phonological, phonetic, and psycholinguistic re- search on prosodic structure.* INTRODUCTION 1. The Hebrew Bible text is annotated with a system of diacritic marks called ACCENTS. These accents, assigned to every word in the Bible, parse each verse in minute detail. The nature and purpose of this complex system of represen- tation, developed in and around Tiberias over several generations up until the tenth century, has been much debated. Some scholars have suggested that it is intended to be a kind of syntactic representation, or else a system devised to indicate logical and semantic relations among words and phrases. I will argue that it is best viewed as providing a prosodic representation, designed to in- dicate the correct phrasing of the text. I will show that many characteristic properties of the Tiberian system of accents follow directly from basic prin- ciples of prosodic theory and are typical of the prosodic structure of other languages. Viewing the Tiberian accentuation from the perspective of prosodic theory also sheds light on elements of the system which have long appeared to be problematic. Conversely, the Tiberian system merits study, as it displays an extraordinary complexity and subtlety which may contribute to our under- standing of prosodic representation in general. The plan of this article is as follows. Section 2 is a brief introduction to the system of accents, its background, and its purpose. In ?3 the accentual rep- resentation is considered in the light of recent proposals in prosodic theory,

Selective document retrieval method and system

Abstract: A method and system for storing and selectively retrieving information, such as words, from a document set. The method includes generating an image data set representative of the information contained in the document set. The method also involves generating a text data set representative of a text portion of the information contained in the document set. A text-image correspondence (TIC) table is generated that includes data representative of coordinates information corresponding to each phrase of the document set. A search phrase is identified in response to user-specified search criteria and the search phrase is identified in the text image data set. Then, the TIC table is used to identify the coordinates information corresponding to the search phrase identified in the text data set. A display of the portion of the page containing the search phrase is generated using the coordinates information.

Hybrid Connectionist Natural Language Processing

Abstract: Introduction. Connectionist and hybrid models for language understanding. A hybrid connectionist scanning understanding of phrases. Structural phrase analysis in a hybrid separated model. Structural phrase analysis in a hybrid integrated model. Contextual phrase analysis in a hybrid separated model. Contextual phrase analysis in a hybrid integrated model. General summary and discussion. Appendix.

Using Non-Sentences: An Application of Relevance Theory

Abstract: Michael Dummett has nicely expressed a rather widespread doctrine about the primacy of sentences. He writes: "you cannot DO anything with a word — cannot effect any conventional (linguistic) act by uttering it — save by uttering some sentence containing that word ...". In this paper we argue that this doctrine is mistaken: it is not only sentences, but also ordinary words and phrases which can be used in isolation. The argument involves two steps. First: we show — using Sperber and Wilson's relevance theory — that an utterance of "John's father" could COMMUNICATE a proposition. Second: we point out that, in this context, this proposition would be asserted rather than merely implicated. Because there is nothing importantly idiosyncratic about the phrase "John's father", we infer that words and phrases generally can be used in isolation to make assertions.

Sentential Context and On-Line Lexical Decision

Abstract: J. L. Nicol and D. Swinney (1989) reported that lexical decision response times to a test word that was related to the implicit object of a verb were faster, when tested immediately after the verb, than response times to a control test word. They concluded from this result that the relation between the implicit object and the verb was understood during comprehension. In G. McKoon, R. Ratcliff, and G. Ward (1994), another interpretation was suggested, that the faster lexical decision response times were due to the better semantic-pragmatic fit of the related test words to the sentential contexts, and experimental data were presented to support this interpretation. In response, J.L. Nicol, J. D. Fodor, and D. Swinney (1994) pointed out some possible problems with these experiments. The experiment presented in this article undermines their arguments. By using sentences in which there was no implicit object, exactly the same pattern of results was found as originally reported by J. L. Nicol and D. Swinney, thereby implicating some factor other than syntactic processing of an implicit object as an explanation of their finding.

The language development of children with Down syndrome: First words to two-word phrases

Abstract: A longitudinal study was conducted to investigate the early language development of children with Down syndrome. Using parental records of their children's speech, the study investigated very early language development, from first words to the consolidation of two-word phrases. Seventeen children were studied, all with Down syndrome and the age range of these subjects when the study began was one year to four years. The mean age for reaching the ten word stage was 27.3 months, a delay of around 12 months. The results showed that the children with Down syndrome studied here learned very similar words to those of typically developing children. The results also indicated that, as with typically developing children, some children with Down syndrome experienced a vocabulary explosion. The mean age for this language explosion was 30 months, with a mean vocabulary of 24.4 words. However, it was clear that there was a wide range of individual differences with some children showing no explosion. The mean age for two-word phrase consolidation was 36.9 months, suggesting a delay of around 18 months. The results indicated that, having reached the ten word stage, children with Down syndrome proceed in their language development at a slower rate to a two-word stage than typically developing children. Detailed individual profiles are included to illustrate the wide range of individual differences observed in the rate of language acquisition in this group.

Lexical government and tone group formation in Xiamen Chinese

Abstract: The study of the relation between syntactic structure and phonological representation has attracted the attention of many phonologists in the past few years. One important contribution to this field of study is Chen's (1987) work on Xiamen Chinese tone sandhi domains. He suggests that the syntax–phonology relation appeals to syntactic information such as category types and the edges of syntactic bracketings. This insight has been further elaborated in the general theory of the syntax—phonology relation of Selkirk (1986). In this theory, the relation between syntactic structure and prosodic structure above the foot and below the intonational phrase is defined in terms of the edges of syntactic constituents of designated types. More precisely, this theory incorporates two hypotheses. One is that there are designated category types in syntactic structure with respect to which one end (Right or Left) of the designated category is relevant in the formulation of a prosodic constituent C, which extends from one instance of the appropriate end (R/L) of the designated category to the next. This hypothesis has been called the End Parameter.

The syntax of questions in child English

Abstract: The purpose of this article is to provide a contemporary Government-and-Binding (GB) reinterpretation and evaluation of Klima & Bellugi's classic 1966 work on the acquisition of interrogatives. I argue that the central insight of K&B's paper can be captured by positing that wh-questions in Child English involve a wh-pronoun positioned in the head complementizer (C) position within the Complementizer Phrase (CP) (so blocking auxiliary inversion if this involves positioning an inverted auxiliary in C) and that in the transition to Adult English, children come to learn that wh-questions involve a wh-phrase superficially positioned in the specifier position within CP. I argue that the wh-in-C analysis poses both developmental problems (in that it fails to account for child structures involving a preposed wh-phrase with an uninverted auxiliary) and potential theoretical problems (in that long movement of a wh-head may violate locality principles). I then consider two alternative accounts of wh-questions which posit that wh-movement involves movement of a wh-phrase from the very earliest stages of development. The first of these is an adjunction account, on which wh-phrases are analysed as clausal adjuncts in Child English (adjoined to the Verb Phrase (VP) in the earliest stages and to the Inflection Phrase (IP) in later stages). I note, however, that this provides no principled account of the absence of auxiliary inversion in child wh-questions, and poses continuity problems (especially within a framework such as that of Cinque (1990) in which it is assumed that wh-phrases never adjoin to VP or IP). Finally, I consider an alternative account on which initial wh-phrases are analysed as occupying the specifier position within CP at all stages of development. I note that the problem posed by this analysis is accounting for the absence of auxiliary inversion in early wh-questions, and offer an account which posits that children overgeneralize specifier-head agreement from IP to CP.

Automatic alignment in parallel corpora

Abstract: This paper addresses the alignment issue in the framework of exploitation of large bimultilingual corpora for translation purposes. A generic alignment scheme is proposed that can meet varying requirements of different applications. Depending on the level at which alignment is sought, appropriate surface linguistic information is invoked coupled with information about possible unit delimiters. Each text unit (sentence, clause or phrase) is represented by the sum of its content tags. The results are then fed into a dynamic programming framework that computes the optimum alignment of units. The proposed scheme has been tested at sentence level on parallel corpora of the CELEX database. The success rate exceeded 99%. The next steps of the work concern the testing of the scheme's efficiency at lower levels endowed with necessary bilingual information about potential delimiters.

Synthesis of speech from text

Abstract: An apparatus for synthesizing speech from text includes a language processing section which determines an accent environment of each mora of the text. In a basic accent pattern table, a basic accent pattern is classified according to the accent environment of the mora. The basic accent pattern includes a pitch data which is edited from real voice data according to the accent environment. A basic accent pattern processing section selects the basic accent pattern of each more from the basic accent pattern table according to the accent environment and processes the basic accent pattern in pitch according to the accent environment. A correcting section receives the corrected pitch data in the basic accent patter processing section and corrects the corrected pitch data according to the number of mora in each phrase and the position of the mora in phrase so as to correct the data into the corrected accent component. A phrase pattern processing section determines a phrase component according to the number of mora in each phrase which is of the accent environment. A speech synthesizing section synthesizes speech according to an accent control pattern of the text which is obtained by adding the accent pattern and the phrase pattern.

Karaoke apparatus sounding instrumental accompaniment and back chorus

Abstract: A karaoke apparatus is comprised of a data supply, a tone generator, an ADPCM decoder and a pitch shifter for sounding a requested karaoke song containing an instrumental accompaniment and a back chorus. The data supply supplies an ADPCM data representative of a phrase back chorus and being compressed by a variable compression condition, and supplies a song data containing accompaniment information prescriptive of an instrumental accompaniment, additional information prescriptive of a common back chorus and decoding information indicative of the variable compression condition of the ADPCM data, and further supplies key information which determines a pitch shift of the karaoke song. The tone generator processes the accompaniment information and the additional information to generate a first audio signal effective to sound synthesized tones of the instrumental accompaniment and the common back chorus. The ADPCM decoder operates in synchronization with the tone generator to decode the ADPCM data according to the decoding information for reproducing a second audio signal effective to sound the phrase back chorus concurrently with the instrumental accompaniment. The pitch shifter processes the second audio signal according to the key information to carry out the pitch shift of the phrase back chorus.

Developing Syntactic Sensitivity in Reading Through Phrase-cued Texts

Abstract: Highlights the benefits of developing students' skills in grouping text into syntactically appropriate units as a way of addressing reading problems and increasing proficiency