Computer-implemented voice response method using a dialog state diagram to facilitate operator intervention
View Patent ↗A voice response system attempts to respond to spoken user input and to provide computer-generated responses. If the system decides it cannot provide valid responses, the current state of user session is determined and forwarded to a human operator for further action. The system maintains a recorded history of the session in the form of a dialog history log. The dialog history and information as to the reliability of past speech recognition efforts is employed in making the current state determination. The system includes formatting rules for controlling the display of information presented to the human operator.
1. A method for providing voice responses to spoken input items received from a user during a session between the user and a voice response system, comprising:
providing computer recognition of spoken input items;
providing system responses to spoken input items;
storing, in a dialog history log, a record of recognized spoken input items and system responses thereto, said record representing a dialog history;
responsive to a determination that the system cannot provide a valid system response to spoken input items, using a dialog state determination model to determine the current state of the dialog with the user based on the dialog history log and a dialog state diagram definition file defining each expected dialog state for the session;
forwarding the determined current dialog state to a visual information display remote from the user for use by a human operator other than the user; and
forwarding a dialog state diagram including a representation of each dialog state defined by the dialog state diagram definition file, including at least one dialog state, other than a transition between dialog states, not yet entered during the session, to the visual information display for use by the human operator.
2. The method according to claim 1 , wherein the dialog history log includes at least one speech recognition result occurring between the start of user input and the time of forwarding of dialog state information to the human operator, a reliability value for the speech recognition result, an already entered item in the speech recognition result, information identifying any system response to the speech recognition result, information indicating the start of the user input, and information indicating the time of forwarding.
3. The method according to claim 1 , wherein the dialog state diagram definition file defines each expected dialog state and the format in which each dialog state and associated input information is to be displayed to the human operator.
4. The method according to claim 1 , wherein the dialog state determination model includes empirical data and/or algorithms for determining a current dialog state from the dialog history log.
5. The method according to claim 1 , wherein the method further comprises steps of:
calculating the reliability of an item entered by the user by taking into account an input history; and
presenting the item on the visual information display in a manner dependent on the calculated reliability.
6. The method according to claim 1 , further including the step of highlighting the current dialog state in the dialog state diagram displayed on the visual information display.