Abstract: A system and method for creating user voice profiles enables a user to create a single user voice profile that is compatible with one or more voice servers. Such a system includes a training server that receives audio information from a client associated with a user and stores the audio information and corresponding textual information. The system further includes a training server adaptor. The training server adaptor is configured to receive a voice profile format and a communication protocol corresponding to one of the plurality of voice servers, convert the audio information and corresponding textual information into a format compatible with the voice profile format and communication protocol corresponding to the one of the plurality of voice servers, and provide the converted audio information and corresponding textual information to the one of the plurality of voice servers.

Abstract: Disclosed are systems and methods for analyzing and improving document readability. For example, a method of reducing writing problems in existing prose is disclosed that can deal with problems related to the word “it” in text. Such a method can include the following steps: scanning prose to determine whether a particular writing problem exists in the prose; determining if at least one sign is present, the sign indicating a possible occurrence or absence of the writing problem, the sign comprising the word “it”; and specifying a proposed edit to a user of the method, the proposed edit changing the prose in a way that addresses the writing problem. Various other rules for improving text are also disclosed.

Abstract: A method, system, and computer program product for national language support. National language support for an application is provided by recording translations of a text string in corresponding different languages in a single property file so as to allow display of the translations in the property file. One of the translations of the text string recorded in the property file is selected for use by an application based on a locale associated with the execution of the application and the selected one of the translations is used in the execution of the application.

Abstract: Mechanisms are disclosed for incorporating prototype information into probabilistic models for automated information processing, mining, and knowledge discovery. Examples of these models include Hidden Markov Models (HMMs), Latent Dirichlet Allocation (LDA) models, and the like. The prototype information injects prior knowledge to such models, thereby rendering them more accurate, effective, and efficient. For instance, in the context of automated word labeling, additional knowledge is encoded into the models by providing a small set of prototypical words for each possible label. The net result is that words in a given corpus are labeled and are therefore in condition to be summarized, identified, classified, clustered, and the like.

Abstract: A system for adding words to an online dictionary used for spellchecking is described. A spellchecker module compares words of an electronic document with words in the online dictionary and identifies a word in the electronic document that is missing from the dictionary. After a user indicates a desire to add the missing word to the dictionary, the spellchecker module determines at least one related-word form of the missing word. The related-word forms depend upon the part of speech of the missing word. The spellchecker can prompt the user to identify the part of speech and then to verify each determined related-word form. The spellchecker concurrently adds the missing word and at least one related-word form of the missing word to the online dictionary in a single ‘add-to-dictionary’ operation.

Abstract: A “STAC Codec” provides lossless audio compression and decompression by processing an audio signal using integer-reversible modulated lapped transforms (MLT) to produce transform coefficients. Transform coefficients are then encoded using a backward-adaptive run-length Golomb-Rice (RLGR) encoder to produce losslessly compressed audio signals. In additional embodiments, further compression gains are achieved via an inter-block spectral estimation and data sorting strategy. Further, compression in the transform domain allows the bitstream to be partially decoded, using the corresponding RLGR decoder, to reconstruct the frequency-domain coefficients. These frequency-domain coefficients are then directly used to speed up various transform-domain based applications such as transcoding media to lossy or other formats, search, identification, visualization, watermarking, etc.

Abstract: On an encoder-side, a multi-channel input signal is analyzed for obtaining smoothing control information, which is to be used by a decoder-side multi-channel synthesis for smoothing quantized transmitted parameters or values derived from the quantized transmitted parameters for providing an improved subjective audio quality in particular for slowly moving point sources and rapidly moving point sources having tonal material such as fast moving sinusoids.

Abstract: Systems and methods for identifying translation pairs from web pages are provided. One disclosed method includes receiving monolingual web page data of a source language, and processing the web page data by detecting the occurrence of a predefined pattern in the web page data, and extracting a plurality of translation pair candidates. Each of the translation pair candidates may include a source language string and target language string. The method may further include determining whether each translation pair candidate is a valid transliteration. The method may also include, for each translation pair that is determined not to be a valid transliteration, determining whether each translation pair candidate is a valid translation. The method may further include adding each translation pair that is determined to be a valid translation or transliteration to a dictionary.

Abstract: Techniques are described for providing relevant information to users (e.g., information that is at least potentially of interest to the users). Relevant information for a user may be automatically determined based on a determined context of the user and/or on a request for that information from the user. For example, voice-based information may be obtained from a user in one or more ways, and then analyzed to identify requests or other indications of information of interest and/or to otherwise determine a context of the user that corresponds to potential information of interest. Relevant information for a user may be provided to the user in various ways, such as via a voice-based response during a telephone call and/or via one or more electronic messages sent to the user (e.g., via emails, instant messages, paging messages, SMS or other text messages, etc.).

Abstract: A system and method for speech file processing which provides users with differentially selectable speech file transcripts which can be sent to one or more other users. The speech files may be voicemail messages from which respective voicemail transcripts are created. The voicemail transcripts are provided in a user selectable format from which users may select non-contiguous portions of the transcript.

Abstract: Techniques are provided for globalizing handling of service management items. The techniques include obtaining a service management item in a language convenient to a first of two or more actors, translating the service management item into a language-neutral format to obtain a language-neutral service management item, applying one or more annotators to the service management item, translating the language-neutral service management item into a language convenient to a second of two or more actors acting on the service management item, and routing the translated service management item to the second of two or more actors. Techniques are also provided for generating a database of service management items in a language-neutral format.

Abstract: Systems and methods are provided for automatically building a native phonetic lexicon for a speech-based application trained to process a native (base) language, wherein the native phonetic lexicon includes native phonetic transcriptions (base forms) for non-native (foreign) words which are automatically derived from non-native phonetic transcriptions of the non-native words.

Abstract: A system and computer program product for national language support. National language support for an application is provided by recording translations of a text string in corresponding different languages in a single property file so as to allow display of the translations in the property file. One of the translations of the text string recorded in the property file is selected for use by an application based on a locale associated with the execution of the application and the selected one of the translations is used in the execution of the application.

Abstract: A wireless communication device is disclosed that accepts recorded audio data from an end-user. The audio data can be in the form of a command requesting user action. Likewise, the audio data can be converted into a text file. The audio data is reduced to a digital file in a format that is supported by the device hardware, such as a .wav, .mp3, .vnf file, or the like. The digital file is sent via secured or unsecured wireless communication to one or more server computers for further processing. In accordance with an important aspect of the invention, the system evaluates the confidence level of the of the speech recognition process. If the confidence level is high, the system automatically builds the application command or creates the text file for transmission to the communication device.

Abstract: Methods and systems for extending keyword searching techniques to syntactically and semantically annotated data are provided. Example embodiments provide a Syntactic Query Engine (“SQE”) that parses, indexes, and stores a data set as an enhanced document index with document terms as well as information pertaining to the grammatical roles of the terms and ontological and other semantic information. In one embodiment, the enhanced document index is a form of term-clause index, that indexes terms and syntactic and semantic annotations at the clause level. The enhanced document index permits the use of a traditional keyword search engine to process relationship queries as well as to process standard document level keyword searches. In one embodiment, the SQE comprises a Query Processor, a Data Set Preprocessor, a Keyword Search Engine, a Data Set Indexer, an Enhanced Natural Language Parser (“ENLP”), a data set repository, and, in some embodiments, a user interface or an application programming interface.

Abstract: A method and device for creating a glossary includes a processor operable for executing computer instructions for identifying, in at least one information source, at least one glossary item identifying a part or a component, determining at least one glossary item form as a canonical form, defining, by using the canonical form, at least one syntactic structure, that includes one of the at least one identified glossary items, for each of at least one semantic classes, and searching a second information source for the at least one syntactic structure of the semantic class.

Abstract: A system and method for generating a video sequence having mouth movements synchronized with speech sounds are disclosed. The system utilizes a database of n-phones as the smallest selectable unit, wherein n is larger than 1 and preferably 3. The system calculates a target cost for each candidate n-phone for a target frame using a phonetic distance, coarticulation parameter, and speech rate. For each n-phone in a target sequence, the system searches for candidate n-phones that are visually similar according to the target cost. The system samples each candidate n-phone to get a same number of frames as in the target sequence and builds a video frame lattice of candidate video frames. The system assigns a joint cost to each pair of adjacent frames and searches the video frame lattice to construct the video sequence by finding the optimal path through the lattice according to the minimum of the sum of the target cost and the joint cost over the sequence.

Abstract: A system for conducting a telephonic speech recognition application includes an automated telephone device for making telephonic contact with a respondent and a speech recognition device which, upon the telephonic contact being made, presents the respondent with at least one introductory prompt for the respondent to reply to; receives a spoken response from the respondent; and performs a speech recognition analysis on the spoken response to determine a capability of the respondent to complete the application. If the speech recognition device, based on the spoken response to the introductory prompt, determines that the respondent is capable of competing the application, the speech recognition device presents at least one application prompt to the respondent.

Abstract: To alleviate degradation of sound quality which may be caused by pre-echoes and bit starvation. An acoustic analyzer analyzes an audio signal to calculate perceptual entropy indicating how many bits are required for quantization. A coded bit count monitor monitors the number of coded bits produced from the audio signal and calculates the number of available bits for the current frame. Based on the combination of the perceptual entropy and the number of available bits, a frame division number determiner determines a division number N for dividing a frame of the audio signal into N blocks. An orthogonal transform processor divides a frame by the determined division number and subjects each divided block of the audio signal to an orthogonal transform process, thereby obtaining orthogonal transform coefficients. A quantizer quantizes the orthogonal transform coefficients on a divided block basis.

Abstract: A system and method for translating data from a source language to a target language is provided wherein machine generated target translation of a source sentence is compared to a database of human generated target sentences. If a matching human generated target sentence is found, the human generated target sentence may be used instead of the machine generated sentence, since the human generated target sentence is more likely to be a well-formed sentence than the machine generated sentence. The system and method does not rely on a translation memory containing pairs of sentences in both source and target languages, and minimizes the reliance on a human translator to correct a translation generated by machine translation.