Part of the Theory and Applications of Natural Language Processing book series (NLP)
Building Simulation Environments from Wizard-of-Oz Data
This Chapter describes a framework for training and testing a RL-based policy in simulation, where the simulated environment is obtained from limited amounts of Wizard-of-Oz (WOZ) data.
KeywordsAutomatic Speech Recognition Reward Function Dialogue System Dialogue Strategy Word Error Rate
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
Unable to display preview. Download preview PDF.
© Springer-Verlag Berlin Heidelberg 2011