International Business Machines Corporation

Steven V DeGennaro

Stephane Herman Maes

David Nahamoo

Liam David Comerford

Ponani S Gopalakrishnan

Daniel Mark Coffman

Edward Adam Epstein

Lalit R Bahl

Robert L Mercer

Peter V DeSouza

A method for conversational computing includes executing code embodying a conversational virtual machine, registering a plurality of input/output resources with a conversational kernel, providing an interface between a plurality of active applications and the conversational kernel processing input/output data, receiving input queries and input events of a multi-modal dialog across a plurality of user interface modalities of the plurality of active applications, generating output messages and output events of the multi-modal dialog in connection with the plurality of active applications, managing, by the conversational kernel, a context stack associated with the plurality of active applications and the multi-modal dialog to transform the input queries into application calls for the plurality of active applications and convert the output messages into speech, wherein the context stack accumulates a context of each of the plurality of active applications.

Conversational computing via conversational virtual machine

A conversational computing system that provides a universal coordinated multi-modal conversational user interface (CUI)across a plurality of conversationally aware applications () (i.e., applications that 'speak' conversational protocols) and conventional applications (). The conversationally aware applications () communicate with a conversational kernel () via conversational application APIs (). The conversational kernelcontrols the dialog across applications and devices (local and networked) on the basis of their registered conversational capabilities and requirements and provides a unified conversational user interface and conversational services and behaviors. The conversational computing system may be built on top of a conventional operating system and APIs () and conventional device hardware (). The conversational kernel () handles all I/O processing and controls conversational engines (). The conversational kernel () converts voice requests into queries and converts outputs and results into spoken messages using conversational engines () and conversational arguments (). The conversational application API () conveys all the information for the conversational kernel () to transform queries into application calls and conversely convert output into speech, appropriately sorted before being provided to the user.

A conversational computing system that provides a universal coordinated multi-modal conversational user interface (CUI) () across a plurality of conversationally aware applications () (i.e., applications that 'speak' conversational protocols) and conventional applications (). The conversationally aware maps, applications () communicate with a conversational kernel () via conversational application APIs (). The conversational kernel () controls the dialog across applications and devices (local and networked) on the basis of their registered conversational capabilities and requirements and provides a unified conversational user interface and conversational services and behaviors. The conversational computing system may be built on top of a conventional operating system and APIs () and conventional device hardware (). The conversational kernel () handles all I/O processing and controls conversational engines (). The conversational kernel () converts voice requests into queries and converts outputs and results into spoken messages using conversational engines () and conversational arguments (). The conversational application API () conveys all the information for the conversational kernel () to transform queries into application calls and conversely convert output into speech, appropriately sorted before being provided to the user.

Speech words are recognized by first recognizing each spectral vector identified by a label (feneme), then identifying the word by matching the string of labels against phones using simplified phone machines based on label and transition probabilities and Merkov chains. In one embodiment, a detailed acoustic match word score is combined with an approximate acoustic match word score to provide a total word score for a subject word. In another embodiment, a polling word score is combined with an acoustic match word score to provide a total word score for a subject word. The acoustic models employed in the acoustic matching may correspond, alternatively, to phonetic elements or to fenemes. Fenemes represent labels generated by an acoustic processor in response to a spoken input. Apparatus and method for determining word scores according to approximate acoustic matching and for determining word scores according to a polling methodology are disclosed.

Growing community of inventors

Pawling, NY, United States of America

Steven V DeGennaro