US 9,812,127 B1
Reactive learning for efficient dialog tree expansion
Julien Perez, Grenoble (FR); and Nicolas Monet, Montbonnot-Saint-Martin (FR)
Assigned to Conduent Business Services, LLC, Dallas, TX (US)
Filed by Conduent Business Services, LLC, Dallas, TX (US)
Filed on Apr. 29, 2016, as Appl. No. 15/142,187.
Int. Cl. G10L 15/22 (2006.01); G06F 17/27 (2006.01); G06F 17/24 (2006.01); G10L 15/06 (2013.01)
CPC G10L 15/22 (2013.01) [G06F 17/241 (2013.01); G06F 17/279 (2013.01); G10L 15/063 (2013.01); G10L 2015/0635 (2013.01)] 18 Claims
OG exemplary drawing
 
1. A method for generating dialogs and learning a dialog policy for a dialog system, comprising:
for each of at least one scenario, in which annotators in a pool of annotators serve as virtual agents and users, generating a respective dialog tree in which each path through the tree corresponds to a dialog and nodes of the tree correspond to turn of a dialog, the generation comprising, with a processor:
a) computing a measure of uncertainty for nodes in the dialog tree, comprising:
for each of a plurality of nodes, computing a conflict coefficient Ci which quantifies the diversity of its child-node set, as a function of:

OG Complex Work Unit Drawing
where

OG Complex Work Unit Drawing
is the maximum size of any subtree from the node j, k is the number of child nodes of node j, ni is the number of child nodes of one of these child nodes, n is the mean number of child nodes of the subtrees that have as root the child nodes of the considered node j;
b) identifying a next node to be annotated, based on the measure of uncertainty,
c) selecting an annotator from the pool to provide an annotation for the next node,
d) receiving an annotation from the selected annotator for the next node, and
e) generating a new node of the dialog tree based on the received annotation;
generating a corpus of dialogs from the dialog tree;
learning a dialog policy based on the corpus of dialogs; and
incorporating the learned dialog policy into a dialog system for conducting a dialog between a virtual agent and a user, in which the learned dialog policy predicts, based on a state of the dialog, a next action to perform, the action being converted, by the dialog system, to a next utterance of the virtual agent.