Abstract: An acoustical-signal processing apparatus includes a feature extracting unit that extracts feature data common to each channel signal which forms a multichannel acoustical signal, based on a composite similarity obtained by combining similarities calculated from each channel signal; and a time-base companding unit that executes time compression and time expansion of the multichannel acoustical signal based on the extracted feature data.

Abstract: To provide a pointer position control method and the like for manipulating a pointer more easily. The user moves the pointer P two-dimensionally and perform click and other operations by using only “voice”—by varying the volume and pitch of produced voice without uttering any specific command. The user moves the pointer P by varying the volume and switches the travel direction of the pointer P by changing the pitch. Also, by stopping to vary the volume, the user can automatically enter a fine adjustment mode in which the user can make fine adjustments. Furthermore, the user can perform a click by stopping to produce voice suddenly and return to normal speech recognition mode by keeping silent.

Abstract: The invention relates to remote access systems and methods using automatic speech recognition to access a computer system. The invention also relates to an intelligent agent resident on the computer system for facilitating remote access to, and receipt of, information on the computer system through speech recognition or text-to-speech read-back. The remote access systems and methods can be used by a user of the computer system while traveling. The user can dial into a server system which is configured to interact with the user by automatic speech recognition and text-to-speech conversion. The server system establishes a connection to an intelligent agent running on the user's remotely located computer system by packet communication over a public network. The intelligent agent sources information on the user's computer system or a network accessible to the computer system, processes the information and transmits it to the server system over the public network.

Abstract: A voice-enabled help desk service is disclosed. The service comprises an automatic speech recognition module for recognizing speech from a user, a spoken language understanding module for understanding the output from the automatic speech recognition module, a dialog management module for generating a response to speech from the user, a natural voices text-to-speech synthesis module for synthesizing speech to generate the response to the user, and a frequently asked questions module. The frequently asked questions module handles frequently asked questions from the user by changing voices and providing predetermined prompts to answer the frequently asked question.

Abstract: Grammatical element prediction is used to predict grammatical elements in text fragments (such as phrases or sentences). In one embodiment, a statistical model, using syntax features, is used to predict grammatical elements.

Abstract: A method and apparatus for automating the acquisition, reconstruction, and generation of knowledgebases of associated ideas and using such knowledgebases in many applications including machine translation of human languages, search and retrieval of unstructured text, or other data, based on concept search, voice recognition, data compression, and artificial intelligence systems.

Abstract: A system and method may be disclosed for facilitating the site-specific customization of automated speech recognition systems by providing a customization client for site-specific individuals to update and modify language model input files and post processor input files. In customizing the input files, the customization client may provide a graphical user interface for facilitating the inclusion of words specific to a particular site. The customization client may also be configured to provide the user with a series of formatting rules for controlling the appearance and format of a document transcribed by an automated speech recognition system.

Abstract: An audio encoder and decoder use architectures and techniques that improve the efficiency of multi-channel audio coding and decoding. The described strategies include various techniques and tools, which can be used in combination or independently. For example, an audio encoder performs a pre-processing multi-channel transform on multi-channel audio data, varying the transform so as to control quality. The encoder groups multiple windows from different channels into one or more tiles and outputs tile configuration information, which allows the encoder to isolate transients that appear in a particular channel with small windows, but use large windows in other channels. Using a variety of techniques, the encoder performs flexible multi-channel transforms that effectively take advantage of inter-channel correlation. An audio decoder performs corresponding processing and decoding. In addition, the decoder performs a post-processing multi-channel transform for any of multiple different purposes.

Abstract: A system, method, and program product for translating text. The invention provides a bidirectional translation corpus that is used to translate phrases from a first language to a second language and vice versa. The bidirectional translation corpus has multiple entries, each having a phrase in the first language and a corresponding phrase in the second language. A source phrase is compared with each entry in the bidirectional translation corpus to determine if it matches one of the entries. If a match is found, the corresponding phrase is used as a translated phrase. Otherwise, the phrase is translated using a translation system.

Abstract: A concept-based back translation system includes a target language semantic parser module, a source language semantic parser module, a bi-directional machine translation module, a relevancy judging module, and a back translation display module.

Abstract: Multiple input modalities are selectively used by a user or process to prune a word graph. Pruning initiates rescoring in order to generate a new word graph with a revised best path.

Abstract: A system and method of controlling the movement of a virtual agent while the agent is listening to a human user during a conversation is disclosed. The method comprises receiving speech data from the user, performing a prosodic analysis of the speech data and controlling the virtual agent movement according to the prosodic analysis.

Abstract: The present invention provides method and apparatus for bilingual word alignment, method and apparatus for training bilingual word alignment model. The method for training bilingual word alignment model, comprising: training a bilingual word alignment model for a first language and a second language, using a bilingual corpus of the first and second languages; training a bilingual word alignment model for the second language and a third language, using a bilingual corpus of the second and third languages; and estimating a bilingual word alignment model for the first language and the third language, based on said bilingual word alignment model for the first and second languages and said bilingual word alignment model for the second and third languages.

Abstract: A scalable two-pass scalable probabilistic latent semantic analysis (PLSA) methodology is disclosed that may perform more efficiently, and in some cases more accurately, than traditional PLSA, especially where large and/or sparse data sets are provided for analysis. The improved methodology can greatly reduce the storage and/or computational costs of training a PLSA model. In the first pass of the two-pass methodology, objects are clustered into groups, and PLSA is performed on the groups instead of the original individual objects. In the second pass, the conditional probability of a latent class, given an object, is obtained. This may be done by extending the training results of the first pass. During the second pass, the most likely latent classes for each object are identified.

Abstract: An enhancement system improves the estimate of noise from a received signal. The system includes a spectrum monitor that divides a portion of the signal at more than one frequency resolution. Adaptation logic derives a noise adaptation factor of the received signal. A plurality of devices tracks the characteristics of an estimated noise in the received signal and modifies multiple noise adaptation rates. Weighting logic applies the modified noise adaptation rates derived from the signal divided at a first frequency resolution to the signal divided at a second frequency resolution.

Abstract: A method of multi-lingual speech recognition can include determining whether characters in a word are in a source list of a language-specific alphabet mapping table for a language, converting each character not in the source list according to a general alphabet mapping table, converting each converted character according to the language-specific alphabet mapping table, verifying that each character in the word is in a character set of the language, removing characters not in the character set of the language, and identifying a pronunciation of the word.

Abstract: A multi-channel audio encoder (10) encodes an N-channel audio signal. A first unit (110) generates a first encoded M-channel signal, e.g. a spatial stereo down-mix, for the N-channel signal (N>M). Down-mixers (115, 116, 117) generate first enhancement data for the signal relative to the N-channel audio signal. A second M-channel signal, such as an artistic stereo mix, is generated for the N-channel signal. A processor (123) then generates second enhancement data for the second M-channel signal relative to the first M-channel signal. A second unit (120) generates an output signal comprising the second M-channel signal, the first enhancement data and the second enhancement data. The generator (123) can dynamically select between generating the second enhancement data as absolute enhancement data or as relative enhancement data relative to the second encoded M-channel signal.

Abstract: Various processes are disclosed for conducting database searches by voice. One such process enables a user to efficiently submit a search query by partially spelling the search query (either on a telephone keypad or via voice utterances) and uttering the full search query. Also disclosed are various processes for generating speech recognition grammars for interpreting utterances of search queries. In one such process, search queries are selected from a search query log for incorporation into a speech recognition grammar. The search query log may include or consist of search queries specified by users without the use of voice.

Abstract: Scalable stereo audio coding and decoding method and apparatus are provided. The scalable stereo audio coding method includes transforming a first channel and a second channel audio samples; quantizing the transformed first channel and a second channel audio samples; and coding the quantized first channel audio samples up to a predetermined transition layer and then interleavingly coding the quantized first and second channel audio samples with increasing a layer index from a layer succeeding the transition layer, until coding for a predetermined plurality of layers is finished.

Abstract: Techniques are presented to determine user-interest sensitive notes. User selected passages, user interest information, condensation transformations and optional meaning distortion constraints are identified. User foci expressed by the selected passages are determined based on the similarity of the elements in the selected passages to elements in the user interest information. User sensitive notes are determined by selectively applying the condensation transformations to the selected passages to preferentially retain user foci while eliding less salient information. Meaning distortions constraints are optionally applied in conjunction with condensation transformations or in creating the condensation transformations in order to reduce the likelihood of distorting the meaning of the passage.