Abstract: A machine translation (MT) system may utilize a phrase-based joint probability model. The model may be used to generate source and target language sentences simultaneously. In an embodiment, the model may learn phrase-to-phrase alignments from word-to-word alignments generated by a word-to-word statistical MT system. The system may utilize the joint probability model for both source-to-target and target-to-source translation applications.

Abstract: A system and method that facilitate modeling unobserved speech dynamics based upon a hidden dynamic speech model in the form of segmental switching state space model that employs model parameters including those describing the unobserved speech dynamics and those describing the relationship between the unobserved speech dynamic vector and the observed acoustic feature vector is provided. The model parameters are modified based, at least in part, upon, a variational learning technique. In accordance with an aspect of the present invention, novel and powerful variational expectation maximization (EM) algorithm(s) for the segmental switching state space models used in speech applications, which are capable of capturing key internal (or hidden) dynamics of natural speech production, are provided. For example, modification of model parameters can be based upon an approximate mixture of Gaussian (MOG) posterior and/or based upon an approximate hidden Markov model (HMM) posterior using a variational technique.

Abstract: A voice interactive system includes an acoustic processing part 11 for performing acoustic signal processing with respect to an input voice signal, a voice recognizing part 12 for recognizing the contents of a voice contained in the voice signal after being subjected to the acoustic signal processing, a voice interacting part 13 for transmitting information to a user by using a voice output or a combination of a voice output and another information transmission unit based on the contents of the voice, and a barge-in control part 14 having a barge-in function of suspending the transmission of information based on an input of the acoustic processing part 11, an output thereof, or an input signal from an external input, in the course of transmission of information, wherein the barge-in control part 14 detects at least one feature value from the input signal from the input or the output of the acoustic processing part 11 or the external input, and determines the effectiveness of the barge-in function based on the

Abstract: A system and method summarizes multimedia stored in a compressed multimedia file partitioned into a sequence of segments, where the content of the multimedia is, for example, video signals, audio signals, text, and binary data. An associated metadata file includes index information and an importance level for each segment. The importance information is continuous over as closed interval. An importance level threshold is selected in the closed interval, and only segments of the multimedia having a particular importance level greater than the importance level threshold are reproduced. The importance level can also be determined for fixed-length windows of multiple segments, or a sliding window. Furthermore, the importance level can be weighted by a factor, such as the audio volume.

Abstract: The present invention relates to an apparatus and method for recognizing biological named entity from biological literature based on united medical language system (UMLS). The apparatus and the method receives metathesaurus from the UMLS, constructs a concept name database, a single name database and a category keyterm database, which are language resources to be used recognize a named entity, receives each concept name stored in the concept name database, extracts features of each of the concept names by using data stored in the single name database and the category keyterm database, constructs a rule database by creating rules used to recognize the named entity and filtering the rules by using the extracted features, receives a biological literature, extracts nouns and noun phrases that are candidate named entities, applies the rules stored in the rule database to the nouns and the noun phrases, and recognizes the named entities.

Abstract: Methods are disclosed for estimating language models such that the conditional likelihood of a class given a word string, which is very well correlated with classification accuracy, is maximized. The methods comprise tuning statistical language model parameters jointly for all classes such that a classifier discriminates between the correct class and the incorrect ones for a given training sentence or utterance. Specific embodiments of the present invention pertain to implementation of the rational function growth transform in the context of a discriminative training technique for n-gram classifiers.

Abstract: A method and device within a speech processing unit (SPU) for reducing scheduling delay between the SPU and a radio network node. Within the SPU, data packets are processed in a plurality of time slots that are subunits of frames. The device receives timing information from the node that identifies a beginning and an ending of processing periods in the node. The timing information is utilized to select a time slot within each frame as a target time slot. The target time slot has a position within each frame such that the scheduling delay between the ending of a processing period in the node and the beginning of the target time slot is minimized. Data packets for a particular channel are assigned to the target time slot to reduce the scheduling delay. The phase of the frame is then adjusted by erasing superfluous data packets.

Abstract: A privacy apparatus adds a privacy sound based on a speaker's own voice into the environment, thereby confusing listeners as to which of the sounds is the real source. This permits disruption of the ability to understand the source speech of the user by eliminating segregation cues that the auditory system uses to interpret speech. The privacy apparatus minimizes segregation cues. The privacy apparatus is relatively quiet and thus easily acceptable in a typical open floor design office space. The privacy apparatus contains an A/D converter that converts the speech into a digital signal, a DSP that converts the digital signal into a privacy signal with pre-recorded speech fragments that are summed so that the speech fragments at least partly overlap one another, a D/A converter that converts the privacy signal into an output signal and one or more loudspeakers from which the output signal is emitted.

Abstract: The present invention provides a method and system for processing an audio signal. According to an exemplary method, an audio signal such as a digital voice signal is received and divided into one or more individual unit cycles. An audio speed conversion operation is enabled by repeating or removing one or more of the individual unit cycles. In particular, repeating one or more of the individual unit cycles decreases audio speed, and removing one or more of the individual unit cycles increases audio speed.

Abstract: The present invention comprises a system and method for implementing a Mandarin Chinese speech recognizer with an optimized phone set, and may include a recognizer configured to compare input speech data to phone strings from a vocabulary dictionary that is implemented according to an optimized Mandarin Chinese phone set. The optimized Mandarin Chinese phone set may be implemented with a phonetic technique to separately include consonantal phones and vocalic phones. For reasons of system efficiency, the optimized Mandarin Chinese phone set may preferably be implemented in a compact manner to include only a minimum required number of consonantal phones and vocalic phones to accurately represent Mandarin Chinese speech during the speech recognition procedure.

Abstract: The present invention comprises a system and method for effectively implementing a Mandarin Chinese speech recognition dictionary, and may include a recognizer configured to compare input speech data to phone strings from a vocabulary dictionary that is implemented according to an optimized Mandarin Chinese phone set. The optimized Mandarin Chinese phone set may efficiently be implemented by utilizing an allophone and phonemic variation technique. In addition, the foregoing vocabulary dictionary may be implemented by utilizing unified dictionary optimization techniques to provide robust and accurate speech recognition. Furthermore, the vocabulary dictionary may be implemented as an optimized dictionary to accurately recognize either Northern Mandarin Chinese speech or Southern Mandarin Chinese speech during the speech recognition procedure.

Abstract: The present invention comprises a system and method for implementing a Cantonese speech recognizer with an optimized phone set, and may include a recognizer configured to compare input speech data to phone strings from a vocabulary dictionary that is implemented according to an optimized Cantonese phone set. The optimized Cantonese phone set may be implemented with a phonetic technique to separately include consonantal phones and vocalic phones. For reasons of system efficiency, the optimized Cantonese phone set may preferably be implemented in a compact manner to include only a minimum required number of consonantal phones and vocalic phones to accurately represent Cantonese speech during the speech recognition procedure.

Abstract: A system for automatically generating a dictionary from full text articles extracts <term, definition> pairs from full text articles and stores the <term, definition> pairs as dictionary entries. The system includes a computer readable corpus having a plurality of documents therein. A pattern processing module (120) and a grammar processing module (125) are provided for extracting <term, definition> pairs from the corpus and storing the <term, definition> pairs in a dictionary database (145). A routing processing module selectively routes sentences in the corpus to at least one of the pattern processing module or grammar processing module. In one embodiment, the routing module is incorporated into the pattern processing module which then selectively routes a portion of the sentences to the grammar processing module.