Method And System For Hybrid Decoding For Enhanced End-User Privacy And Low Latency

• US 20190214014A1
• Filed: 11/26/2018
• Published: 07/11/2019
• Est. Priority Date: 05/26/2016
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Quillen, Carl Benjamin, Parihar, Naveen
• Granted Patent: US 10,803,871 B2
• Time in Patent Office: Days
• CPC Class Codes: G10L 15/14 using statistical models, e.g. Hidden Markov Models [HMMs] G10L15/18 takes precedence
• CPC Class Codes: G10L 15/18 using natural language modelling
• CPC Class Codes: G10L 15/197 Probabilistic grammars, e.g. word n-grams
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
• CPC Class Codes: G10L 15/32 Multiple recognisers used in sequence or in parallel; Score combination systems therefor, e.g. voting systems
• CPC Class Codes: H04L 63/04 for providing a confidential data exchange among entities communicating through data packet networks
• CPC Class Codes: H04L 63/0428 wherein the data content is protected, e.g. by encrypting or encapsulating the payload

Abstract

Methods described herein provide functionality for automatic speech recognition (ASR). One such embodiment performs speech recognition using received speech recognition result candidates, where the received candidates were generated by performing Statistical Language Model (SLM) based speech recognition on one or more frames of audio data. In turn, such an embodiment transmits results of the speech recognition, performed using the received speech recognition result candidates, to a user device via a communications network. Results of the speech recognition are available with lower latency than pure cloud based ASR solution.

Claims

1. A computer-implemented method of performing automatic speech recognition, the method implemented by a processor executing program instructions stored in memory and comprising:

receiving speech recognition result candidates from a user device, the received speech recognition result candidates generated by performing speech recognition on one or more frames of audio data on the user device, the speech recognition on the one or more frames of audio data combining two parallel layers of speech recognition that utilize language models capable of being stored on the user device, wherein the two parallel layers include a command-oriented finite state (FST) recognizer layer and a Statistical Language Model (SLM) based layer;

performing speech recognition using the received recognition result candidates; and

transmitting to a user device, via a communications network, results of the speech recognition performed using the received speech recognition result candidates.