skip to main content skip to footer

Computer-Implemented Systems and Methods for a Crowd Source-Bootstrapped Spoken Dialog System

Ramanarayanan, Vikram; Suendermann-Oeft, David; Lange, Patrick; Ivanov, Alexei V.; Evanini, Keelan; Qian, Yao; Yu, Zhou
Patent Issued:
Mar 31, 2020
Patent Number:
ETS Patent
Document Type:
Family ID:
Subject/Key Words:
Patent, Active Patent, Spoken Dialog System (SDS), Crowdsourcing, Bootstrap (Statistics), Video Evaluation, Video, Multimodal Presentation Assessment, Audio-Lingual Methods, Automatic Speech Recognition, Computer-Implemented


Systems and methods are provided for implementing an educational dialog system. An initial task model is accessed that identifies a plurality of dialog states associated with a task, a language model configured to identify a response meaning associated with a received response, and alanguage understanding model configured to select a next dialog statebased on the identified response meaning. The task is provided to aplurality of persons for training. The task model is updated by revisingthe language model and the language understanding model based onresponses received to prompts of the provided task, and the updated taskis provided to a student for development of speaking capabilities.

Read More