Abstract: Described is a technology by which a communications identifier such as a telephone number is used in selecting relevant content (e.g., an advertisement) for outputting as audible and/or visible information to a destination device. The communications identifier is used like a search term/keyword to determine content relevant to that identifier. A telephone number may be the identifier, selected based on a partially-dialed or actually dialed number, or a user's current scrolling position among listed telephone numbers. A telephone number of an incoming call may be used as the identifier. Profile data and/or state data may be used in selecting relevant content. Some of the profile data may be mined based on previous user activities. In one example, the content may be provided via a remote source of the content, and used immediately or cached for later use (e.g., output).

Abstract: Described is a technology by which a user's telephone-related data is aggregated from various sources for use in assisting the user with making telephone calls. For example, call history data corresponding to a landline telephone, a mobile telephone and/or an office telephone of the user may be combined. Other sources include a landline telephone service, a mobile telephone service, an enterprise telephone system or server, a computing device, voice mail data, web page data, electronic message content, instant message content, a contacts list, and/or an information data store. The telephone-related data can be processed (e.g., based on frequency and calling patterns) to determine corresponding probability data to help determine a user's intent in locating a particular recipient to call. The user may access the telephone-related data via voice commands input at one of the user's telephones, or by receiving a visible list of at least part of the telephone-related data.

Abstract: An auto-answer feature is implemented in SIP by configuring a receiving device to automatically acknowledge and answer an incoming call or session from a specific trusted third party. The receiving device may skip to an OK response to an INVITE request when the call is routed through the trusted third party. When the device can automatically answer the incoming call, advanced features such as Push To Talk, Information Tone, Click to Call, and Remote Monitoring may be easily implemented.

Abstract: A push-to-talk feature on a mobile handset is initiated by speaking a recipient's name as the first part of an initial message. A speech recognition device located in the handset or in a push-to-talk server may recognize the recipient's name, determine the proper addressing for the message, establish a push-to-talk session, and deliver the message to the intended recipient. The session may continue until a session timeout has occurred, until another session is started, or until the user otherwise terminates the session.

Abstract: A user interface for a telephone system may include a telephone portion operable to provide telephone administration functionality for each of a plurality of telephones of a telephone system. The telephone portion may include a telephone icon corresponding to at least one telephone of the plurality of telephones, the telephone icon operable to provide the telephone administration functionality for the at least one telephone. Other portions of the user interface may include an external line portion, an extension portion, or a user portion, which may include, respectively, an external line icon, an extension icon, and a user icon, where each icon may provide associated administration functionality. Accordingly, the user interface allows an administrator of the telephone system to easily view and administer the devices and elements of the telephone system.

Abstract: Inbound calls to an enterprise are handled by a call-handling component, by identifying the calling party from intrinsic information about the call, by identifying a potential recipient from non-intrinsic information about the calling party, and by automatically routing the inbound call to a potential recipient.

Abstract: A method and apparatus are provided for reducing noise in a training signal and/or test signal. The noise reduction technique uses a stereo signal formed of two channel signals, each channel containing the same pattern signal. One of the channel signals is “clean” and the other includes additive noise. Using feature vectors from these channel signals, a collection of noise correction and scaling vectors is determined. When a feature vector of a noisy pattern signal is later received, it is multiplied by the best scaling vector for that feature vector and the best correction vector is added to the product to produce a noise reduced feature vector. Under one embodiment, the best scaling and correction vectors are identified by choosing an optimal mixture component for the noisy feature vector. The optimal mixture component being selected based on a distribution of noisy channel feature vectors associated with each mixture component.

Abstract: Call handling based on recipient identification is disclosed. Various embodiments, obtain, process and use recipient identification differently. Recipient identification includes obtaining data, information or vocalizations from a caller that indicate the person to whom the call in intended. One or more announcements relative to the call can be provided based on the data, information or vocalizations from the caller.

Abstract: A speech signal is decoded by determining a production-related value for a current state based on an optimal production-related value at the end of a preceding state, the optimal production-related value being selected from a set of continuous values. The production-related value is used to determine a likelihood of a phone being represented by a set of observation vectors that are aligned with a path between the preceding state and the current state. The likelihood of the phone is combined with a score from the preceding state to determine a score for the current state, the score from the preceding state being associated with a discrete class of production-related values wherein the class matches the class of the optimal production-related value.

Abstract: Portable computing devices automatically interface with other computing devices to interact in a collaborative effort toward providing a single, seamless computing experience for a user. As a user walks into a room with a cellular telephone certain functionality and data can be automatically unloaded to a desktop computer or other device based on a user or device identification or state. For example, a conversation on a cellular telephone can be automatically migrated to a desktop telephone as a user sits down. As a user is about to leave a room for a meeting, the desktop computer can update the telephone with latest versions of certain files. Thus, devices can automatically aggregate and/or decouple to provide a user with a single computing experience. These portable devices can broadcast an extensible set of services to other devices as well as to a host computer or server.

Abstract: A mobile computing device, adapted to be held in the palm of a user's hand, includes an antenna for transmission of information from the mobile computing device. A first microphone, adapted to convert audible speech from the user into speech signals, is positioned at a distal end of the antenna. The antenna is rotatable, while the mobile computing device is held by the user, into a position which directs the first microphone toward the mouth of the user. A speech sensor outputs a sensor signal indicative of whether the user is speaking.

Abstract: A method and apparatus for speaker recognition is provided that matches the noise in training data to noise in testing data using spectral addition. Under spectral addition, the mean and variance for a plurality of frequency components are adjusted in the training data and the test data so that each mean and variance is matched in a resulting matched training signal and matched test signal. The adjustments made to the training data and test data add to the mean and variance of the training data and test data instead of subtracting from the mean and variance.

Abstract: A method and apparatus for speaker recognition is provided that matches the noise in training data to noise in testing data using spectral addition. Under spectral addition, the mean and variance for a plurality of frequency components are adjusted in the training data and the test data so that each mean and variance is matched in a resulting matched training signal and matched test signal. The adjustments made to the training data and test data add to the mean and variance of the training data and test data instead of subtracting from the mean and variance.

Abstract: The present invention relates to methods and systems for handling interactions between a user and a computer. In particular, the present invention relates to methods and systems for handling communication messages from different types of communication interfaces.

Abstract: A method for creating a language model from a task-independent corpus is provided. In one embodiment, a task dependent unified language model is created. The unified language model includes a plurality of context-free grammars having non-terminals and a hybrid N-gram model having at least some of the same non-terminals embedded therein.

Abstract: Methods and systems for recognizing a spoken alias are disclosed. The present invention includes generating a plurality of alias variations based on a discoverable name and creating a phonetic representation for each of the alias variations. The present invention also includes capturing a phonetic pronunciation of the spoken alias. At least one of the created alias variations that has a phonetic representation that corresponds to the captured phonetic pronunciation is selected.

Abstract: A method of speech recognition is provided that identifies a production-related dynamics value by performing a linear interpolation between a production-related dynamics value at a previous time and a production-related target using a time-dependent interpolation weight. The hidden production-related dynamics value is used to compute a predicted value that is compared to an observed value of acoustics to determine the likelihood of the observed acoustics given a sequence of hidden phonological units. In some embodiments, the production-related dynamics value at the previous time is selected from a set of continuous values. In addition, the likelihood of the observed acoustics given a sequence of hidden phonological units is combined with a score associated with a discrete class of production-related dynamic values at the previous time to determine a score for a current phonological state.

Abstract: A speech signal is decoded by determining a production-related value for a current state based on an optimal production-related value at the end of a preceding state, the optimal production-related value being selected from a set of continuous values. The production-related value is used to determine a likelihood of a phone being represented by a set of observation vectors that are aligned with a path between the preceding state and the current state. The likelihood of the phone is combined with a score from the preceding state to determine a score for the current state, the score from the preceding state being associated with a discrete class of production-related values wherein the class matches the class of the optimal production-related value.

Abstract: A method and apparatus determine a channel response for an alternative sensor using an alternative sensor signal and an air conduction microphone signal. The channel response is then used to estimate a clean speech value using at least a portion of the alternative sensor signal.

Abstract: A method and apparatus are provided for reducing noise in a training signal and/or test signal. The noise reduction technique uses a stereo signal formed of two channel signals, each channel containing the same pattern signal. One of the channel signals is “clean” and the other includes additive noise. Using feature vectors from these channel signals, a collection of noise correction and scaling vectors is determined. When a feature vector of a noisy pattern signal is later received, it is multiplied by the best scaling vector for that feature vector and the best correction vector is added to the product to produce a noise reduced feature vector. Under one embodiment, the best scaling and correction vectors are identified by choosing an optimal mixture component for the noisy feature vector. The optimal mixture component being selected based on a distribution of noisy channel feature vectors associated with each mixture component.