Inventor

RASTROW ARIYA

US49 patents

Patents

49 patents

US10811013B1Oct 20, 2020

RASTROW ARIYA

Patents

Intent-specific automatic speech recognition result generation

Lattice encoding using recurrent neural networks

Language model adaptation

Generation of automated message responses

Error tolerant neural network model compression

Speech processing with learned representation of user interaction history

On-device learning in a hybrid speech processing system

Customized speech processing language models

Disambiguation in automatic speech processing

Automatic speech recognition incorporating word usage information

Generative modeling of speech using neural networks

Domain specific endpointing

Compressed finite state transducers for automatic speech recognition

Markov-based sequence tagging using neural networks

Speech recognition using dialog history

Creation of language models for speech recognition

Device-directed utterance detection

Language model speech endpointing

Updating models with trained model update objects

Lattice decoding and result confirmation using recurrent neural networks

Dynamic arc weights in speech recognition models

Generation of predictive natural language processing models

Intent-specific automatic speech recognition result generation

Language and grammar model adaptation

Speech recognition with combined grammar and statistical language models

Class-based discriminative training of speech models

Dialog management for multiple users

Intent-specific automatic speech recognition result generation

Generation of predictive natural language processing models

Endpointing in speech processing

Dialog management for multiple users

Compressed finite state transducers for automatic speech recognition

Predicting pronunciation in speech recognition

Compact HCLG FST

Intermediate data for inter-device speech processing

Entity language models for speech processing

Speech processing techniques

Continuous learning for machine learning models

Device-directed utterance detection

On-device learning in a hybrid speech processing system

Language model adaptation

Generation of automated message responses

Speech processing techniques

Language and grammar model adaptation using model weight data

Automatic speech recognition using language model-generated context

Speech recognition using dialog history

Intermediate data for inter-device speech processing

Low latency audio processing techniques

Context-based automatic speech recognition processing