Google Inc.

Brian E Roark

Bhuvana Ramabhadran

Bryan Christopher Horling

Maryam Garrett

Wan Fen Nicole Quah

Richard Sproat

Jesse Emond

Pedro Jose Moreno Mengibar

Fadi Biadsy

Min Ma

Arindrima Datta

A method includes obtaining a plurality of training data sets each associated with a respective native language and includes a plurality of respective training data samples. For each respective training data sample of each training data set in the respective native language, the method includes transliterating the corresponding transcription in the respective native script into corresponding transliterated text representing the respective native language of the corresponding audio in a target script and associating the corresponding transliterated text in the target script with the corresponding audio in the respective native language to generate a respective normalized training data sample. The method also includes training, using the normalized training data samples, a multilingual end-to-end speech recognition model to predict speech recognition results in the target script for corresponding speech utterances spoken in any of the different native languages associated with the plurality of training data sets.

Language-agnostic multilingual modeling using effective script normalization

Methods, systems, and apparatus, including computer programs stored on a computer-readable storage medium, for transliteration for speech recognition training and scoring. In some implementations, language examples are accessed, some of which include words in a first script and words in one or more other scripts. At least portions of some of the language examples are transliterated to the first script to generate a training data set. A language model is generated based on occurrences of the different sequences of words in the training data set in the first script. The language model is used to perform speech recognition for an utterance.

Transliteration for speech recognition training and scoring

Some implementations are directed to translating chatspeak to a normalized form, where the chatspeak is included in natural language input formulated by a user via a user interface input device of a computing device&#x2014;such as input provided by the user to an automated assistant. The normalized form of the chatspeak may be utilized by the automated assistant in determining reply content that is responsive to the natural language input, and that reply content may be presented to the user via one or more user interface output devices of the computing device of the user. Some implementations are additionally and/or alternatively directed to providing, for presentation to a user, natural language output that includes chatspeak in lieu of a normalized form of the chatspeak, based at least in part on a 'chatspeak measure' that is determined based on past usage of chatspeak by the user and/or by additional users.

Generating output for presentation in response to user interface input, where the input and/or the output include chatspeak

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, relating to enhanced maximum entropy models. In some implementations, data indicating a candidate transcription for an utterance and a particular context for the utterance are received. A maximum entropy language model is obtained. Feature values are determined for n-gram features and backoff features of the maximum entropy language model. The feature values are input to the maximum entropy language model, and an output is received from the maximum entropy language model. A transcription for the utterance is selected from among a plurality of candidate transcriptions based on the output from the maximum entropy language model. The selected transcription is provided to a client device.

Enhanced maximum entropy models

Growing community of inventors

Portland, OR, United States of America

Brian E Roark