Google Inc.

Daisy Stanton

Sean Matthew Shannon

Thomas Edward Bagby

Eric Dean Battenberg

Soroosh Mariooryad

Russell John Wyatt Skerry-Ryan

David Teh-hwa Kao

Adam Dingle

Susannah Raub

Jun Yin

Vijay Sainath Thadkal

Jeffrey Jar-hou Chin

Russell John-Wyatt Skerry-Ryan

David Teh-Hwa Kao

Systems and methods for text-to-speech with novel speakers can obtain text data and output audio data. The input text data may be input along with one or more speaker preferences. The speaker preferences can include speaker characteristics. The speaker preferences can be processed by a machine-learned model conditioned on a learned prior distribution to determine a speaker embedding. The speaker embedding can then be processed with the text data to generate an output that includes audio data descriptive of the text data spoken by a novel speaker.

Neural-network-based text-to-speech model for novel speaker generation

A method for estimating an embedding capacity includes receiving, at a deterministic reference encoder, a reference audio signal, and determining a reference embedding corresponding to the reference audio signal, the reference embedding having a corresponding embedding dimensionality. The method also includes measuring a first reconstruction loss as a function of the corresponding embedding dimensionality of the reference embedding and obtaining a variational embedding from a variational posterior. The variational embedding has a corresponding embedding dimensionality and a specified capacity. The method also includes measuring a second reconstruction loss as a function of the corresponding embedding dimensionality of the variational embedding and estimating a capacity of the reference embedding by comparing the first measured reconstruction loss for the reference embedding relative to the second measured reconstruction loss for the variational embedding having the specified capacity.

Variational embedding capacity in expressive end-to-end speech synthesis

A system for generating an output audio signal includes a context encoder, a text-prediction network, and a text-to-speech (TTS) model. The context encoder is configured to receive one or more context features associated with current input text and process the one or more context features to generate a context embedding associated with the current input text. The text-prediction network is configured to process the current input text and the context embedding to predict, as output, a style embedding for the current input text. The style embedding specifies a specific prosody and/or style for synthesizing the current input text into expressive speech. The TTS model is configured to process the current input text and the style embedding to generate an output audio signal of expressive speech of the current input text. The output audio signal has the specific prosody and/or style specified by the style embedding.

Controlling expressivity in end-to-end speech synthesis systems

A computer-implemented method includes receiving, at a translation server in communication with a network, a request for a translation of text in a source language to a target language. At least a portion of the text is translated at the translation server from the source language to the target language to obtain a translated version of the text in the target language. Translating the text includes determining one or more terms from the text corresponding to a pre-defined term translator, applying the pre-defined term translator to the one or more terms, and translating a set of additional terms from the text from the source language to the target language via a translation model. The method additionally includes providing, via the translation server, the translated version of the text to a web server.

Document translation including pre-defined term translator and translation model

A system for searching an object environment includes harvesting and indexing applications to create a search database and one or more indexes into the database. A scoring application determines the relevance of the objects, and a querying application locates objects in the database according to a search term. One or more of the indexes may be implemented by a hash table or other suitable data structure, where algorithms provide for adding objects to the indexes and searching for objects in the indexes. A ranking scheme sorts searchable items according to an estimate of the frequency that the items will be used in the future. Multiple indexes enable a combined prefix title and full-text content search of the database, accessible from a single search interface.

Combined title prefix and full-word content searching

Temporal ranking scheme for desktop searching

Growing community of inventors

Mountain View, CA, United States of America

Daisy Stanton