Abstract: A method and system are provided by which a wireless mobile device takes a vocally entered query and transmits it in a text message format over a wireless network to a search engine; receives search results based on the query from the search engine over the wireless network; and displays the search results.

Abstract: A method and system are provided by which a wireless mobile device takes a vocally entered query and transmits it in a text message format over a wireless network to a search engine; receives search results based on the query from the search engine over the wireless network; and displays the search results.

Abstract: The invention is a method of improving the performance of a speech recognizer. The method generally involves: providing a lexicon for the speech recognizer; monitoring a user's interaction with a network; accessing a plurality of words associated with the monitored interaction; and including the plurality of words in the lexicon.

Abstract: A method of operating a mobile communication device, the method involving: over a wireless messaging channel receiving a text message that contains a non-text representation of an utterance; extracting the non-text representation from the text message; synthesizing an audio representation of the spoken utterance from the non-text representation; and playing the synthesized audio representation through an audio output device on the mobile communication device.

Abstract: A method for speech recognition. The method uses a single pronunciation estimator to train acoustic phoneme models and recognize utterances from multiple languages. The method includes accepting text spellings of training words in a plurality of sets of training words, each set corresponding to a different one of a plurality of languages. The method also includes, for each of the sets of training words in the plurality, receiving pronunciations for the training words in the set, the pronunciations being characteristic of native speakers of the language of the set, the pronunciations also being in terms of subword units at least some of which are common to two or more of the languages. The method also includes training a single pronunciation estimator using data comprising the text spellings and the pronunciations of the training words.

Abstract: A method for automatic speech recognition includes determining for an input signal a plurality scores representative of certainties that the input signal is associated with corresponding states of a speech recognition model, using the speech recognition model and the determined scores to compute an average signal, computing a difference value representative of a difference between the input signal and the average signal, and processing the input signal in accordance with the difference value.

Abstract: A mobile voice communication device including a wireless transceiver circuit for transmitting and receiving voice communications and for receiving data; a digital processor; and a memory storing application program code which when executed on the digital processor causes the mobile voice communication device to provide predetermined functionality to the user of the mobile voice communication device. The predetermined functionality includes basic features and it includes enhanced features that are in addition to the basic features. The application program code has a deactivated state in which the mobile voice communication device provides the basic features to the user without providing the enhanced features and an activated state in which the mobile voice communication device provides the enhanced features.

Abstract: One aspect of the invention involves word recognition that uses scrollable choice lists in which choices are listed in character-order. Another aspect relates to a scrollable, visually-displayed word recognition choice list, where the recognition candidates on the choice list are each associated with a choice-selecting symbol the user can use to select a desired recognition candidate by pressing an associated button, and where the same choice-selecting symbol is used for different choices displayed on the display at different times as a result of scrolling. Another aspect of the invention relates to providing a choice list of best scoring characters for a particular character position in the spelling of a filter that is used to filter word recognition. Another aspect of the invention relates to a choice list used in word recognition in which the choice list can be scrolled horizontally.

Abstract: Channel normalization for automatic speech recognition is provided. Statistics are measured from an initial portion of a speech utterance. Feature normalization parameters are estimated based on the measured statistics and a statistically derived mapping relating measured statistics and feature normalization parameters. In some examples, the measured statistics comprise measures of an energy from the initial portion of the speech utterance. In some examples, measures of the energy comprise extreme values of the energy.

Abstract: Large vocabulary speech recognition can automatically turn recognition off in one or more ways. A user command can turn on recognition that is automatically turned off after the next end of utterance. A plurality of buttons can each be associated with a different speech mode and the touch of a given button can turn on, and then automatically turn off, the given button's associated speech recognition mode. These selectable modes can include large vocabulary and alphabetic entry modes, or continuous and discrete modes. A first user input can start recognition that allows a sequence of vocabulary words to be recognized and a second user input can start recognition that turns off after one word has been recognized. A first user input can start recognition that allows a sequence of utterances to be recognized and a second user input can start recognition that allows only a single utterance to be recognized.

Abstract: A method for speech recognition. The method uses a single pronunciation estimator to train acoustic phoneme models and recognize utterances from multiple languages. The method includes accepting text spellings of training words in a plurality of sets of training words, each set corresponding to a different one of a plurality of languages. The method also includes, for each of the sets of training words in the plurality, receiving pronunciations for the training words in the set, the pronunciations being characteristic of native speakers of the language of the set, the pronunciations also being in terms of subword units at least some of which are common to two or more of the languages. The method also includes training a single pronunciation estimator using data comprising the text spellings and the pronunciations of the training words.

Abstract: A method of constructing a list of alternate transcripts from a recognized transcript includes generating a list of close call records, matching partial sub-histories from the recognized transcript with one of the history pairs stored in each of the records, and substituting the other of the history pairs for the partial sub-history of the recognized transcript. A close call record is generated each time a pair of partial hypotheses attempt to seed a common word. Each close call record includes history information and scoring information associated with a particular pair of partial hypotheses seeding a common word. Alternate transcripts are constructed by substituting close call histories for partial histories of the recognized transcripts, and also by substituting close call histories for partial histories of other alternate transcript.

Abstract: Word recognition enables a user to have a selected transformation performed on a given word produced by word recognition. In one aspect of the invention, a selectable transformation changes the given word to a differently spelled word having the same word root. In another, a selectable transformation changes a given word to one or more of its homonyms. In yet another, a selectable transformation changes the given word between a representation that spells the word with letters and one that does not. In one aspect of the invention a user can select to display a choice list of transformed words corresponding to a given recognized word and then select to have one of the listed transformed words replace the given word. In another aspect of the invention word recognition favors recognition of words corresponding to a user selected part of speech.

Abstract: Text-to-speech (TTS) generation is used in conjunction with large vocabulary speech recognition to say words selected by the speech recognition. The software for performing the large vocabulary speech recognition can share speech modeling data with the TTS software. TTS or recorded audio can be used to automatically say both recognized text and the names of recognized commands after their recognition. The TTS can automatically repeats text recognized by the speech recognition after each of a succession of end of utterance detections. A user can move a cursor back or forward in recognized text, and the TTS can speak one or more words at the cursor location after each such move. The speech recognition can be used to produces a choice list of possible recognition candidates and the TTS can be used to provide spoken output of one or more of the candidates on the choice list.

Abstract: A method and a system for testing a voice enabled application on a target device, the method including conducting one or more interactions with the target device, at least some of the interactions including presenting an acoustic utterance in an acoustic environment to the target device, receiving an output of the target device in response to the acoustic utterance, and comparing the output to an output expected from the acoustic utterance.

Abstract: Alphabetic filtering of the speech recognition of words uses a key press to indicate a desired character in an alphabetic filter string, where each key press represents two or more letters. The key presses can be disambiguated by recognizing a key-disambiguation utterance in association with a given key press. A user can select a desired recognition candidate from a choice list produced by such filtered word recognition. Ambiguous alphabetic filtering can be performed iteratively in response to the addition of successive ambiguous key presses. A user can select to re-recognize the utterance using filtering based on ambiguous key input after seeing the results of recognition without such filtering. Unambiguous alphabetic filtering can be performed by using multiple presses of an ambiguous key to disambiguate which letter is intended. A user can select between entering text by either large vocabulary speech recognition or by spelling text by pressing phone keys.

Abstract: A method of sending a voice message via a mobile communication device, said method involving: receiving an utterance from a user of the mobile communication device; generating a non-text representation of the received utterance; inserting the non-text representation into a body of a text message; and sending the text message over a wireless messaging channel from the mobile communication device to a recipient's device.

Abstract: A method implemented on a mobile device that includes speech recognition functionality involves presenting to a user of the mobile device a voice-control interface that supports two types of commands at a common level of the interface, the two types of commands including a first type and a second type, the first type being command and control commands and the second type being search request commands. The method further involves: receiving an utterance from the user that corresponds to a command of either of the first type or the second type; recognizing the utterance; if the received utterance is a command of the first type, performing a corresponding command and control function; and if the received utterance is a command of the second type, generating a representation of a corresponding search request and then using the representation to request a search that is responsive to the search request.

Abstract: Methods and systems for providing voice-mediated search capability to a mobile communications device involve receiving a signal from the mobile device that includes a representation of a spoken search request from a user of the mobile device, using speech recognition software to convert the search request into a text search request, extracting side information contained implicitly within the received signal, using the extracted side information to assign the user to a category, sending the text search request and the user category to content providers, receiving from the content providers content that is responsive to the text search request and the user category, and sending to the mobile device search results that are based on content from content providers. The methods and systems further involve sending searches and user categories to advertising providers, and sending advertisements returned by the advertising providers to the mobile device along with the search results.

Abstract: A method implemented on a mobile device that includes speech recognition functionality involves: receiving an utterance from a user of the mobile device, the utterance including a spoken search request; recognizing that the utterance includes a spoken search request; sending a representation of the spoken search request to a remote server over a wireless data connection; receiving search results over the wireless data connection that are responsive to the search request; storing the results on the mobile device; receiving a subsequent search request; performing a subsequent search responsive to the subsequent search request to generate subsequent search results, the subsequent search including searching the stored search results; and presenting the subsequent results on the mobile device.