Microsoft Technology Licensing, LLC

Advanced Technology & Materials Co., Ltd

Pei Zhao

Bo Yan

Kaisheng N Yao

Max Leung

Yu Shi

Mei-Yuh Hwang

Jian Luan

Malone Ma

Jian Wang

Lei He

Fileno A Alleva

Sheng Zhao

Yiu-Ming Leung

Guodong Liu

Quan Li

Zhiying Zhang

Lidong Ding

Shaoxiong Zhou

Zhe Geng

Wenzhi Chen

An example intent-recognition system comprises a processor and memory storing instructions. The instructions cause the processor to receive speech input comprising spoken words. The instructions cause the processor to generate text results based on the speech input and generate acoustic feature annotations based on the speech input. The instructions also cause the processor to apply an intent model to the text result and the acoustic feature annotations to recognize an intent based on the speech input. An example system for adapting an emotional text-to-speech model comprises a processor and memory. The memory stores instructions that cause the processor to receive training examples comprising speech input and receive labelling data comprising emotion information associated with the speech input. The instructions also cause the processor to extract audio signal vectors from the training examples and generate an emotion-adapted voice font model based on the audio signal vectors and the labelling data.

Intent recognition and emotional text-to-speech learning

The technology relates to converting text to speech utilizing recurrent neural networks (RNNs). The recurrent neural networks may be implemented as multiple modules for determining properties of the text. In embodiments, a part-of-speech RNN module, letter-to-sound RNN module, a linguistic prosody tagger RNN module, and a context awareness and semantic mining RNN module may all be utilized. The properties from the RNN modules are processed by a hyper-structure RNN module that determine the phonetic properties of the input text based on the outputs of the other RNN modules. The hyper-structure RNN module may generate a generation sequence that is capable of being converting to audible speech by a speech synthesizer. The generation sequence may also be optimized by a global optimization module prior to being synthesized into audible speech.

Hyper-structure recurrent neural networks for text-to-speech

An example text-to-speech learning system performs a method for generating a pronunciation sequence conversion model. The method includes generating a first pronunciation sequence from a speech input of a training pair and generating a second pronunciation sequence from a text input of the training pair. The method also includes determining a pronunciation sequence difference between the first pronunciation sequence and the second pronunciation sequence; and generating a pronunciation sequence conversion model based on the pronunciation sequence difference. An example speech recognition learning system performs a method for generating a pronunciation sequence conversion model. The method includes extracting an audio signal vector from a speech input and applying an audio signal conversion model to the audio signal vector to generate a converted audio signal vector. The method also includes adapting an acoustic model based on the converted audio signal vector to generate an adapted acoustic model.

Speech recognition and text-to-speech learning system

Pronunciation issues for synthesized speech are automatically detected using human recordings as a reference within a Speech Recognition Assisted Evaluation (SRAE) framework including a Text-To-Speech flow and a Speech Recognition (SR) flow. A pronunciation issue detector evaluates results obtained at multiple levels of the TTS flow and the SR flow (e.g. phone, word, and signal level) by using the corresponding human recordings as the reference for the synthesized speech, and outputs possible pronunciation issues. A signal level may be used to determine similarities/differences between the recordings and the TTS output. A model level checker may provide results to the pronunciation issue detector to check the similarities of the TTS and the SR phone set including mapping relations. Results from a comparison of the SR output and the recordings may also be evaluation by the pronunciation issue detector. The pronunciation issue detector outputs a list that lists potential pronunciation issue candidates.

Speech recognition assisted evaluation on text-to-speech pronunciation issue detection

The invention belongs to the technical field of rapid solidification of amorphous alloy and concretely relates to an iron-based amorphous alloy broad ribbon, wherein the width is 220-1000 mm, the thickness is 0.02-0.03 mm, the transversal thickness deviation is smaller than +/&#x2212;0.002 mm, the lamination factor is larger than 0.84, the saturation magnetic-flux density is larger than 1.5 T, the iron loss is smaller than 0.20 W/kg under the conditions that the frequency is 50 Hz and the maximum magnetic-flux density is 1.3 T, and the exciting power is smaller than 0.50 VA/kg. The invention also relates to a manufacturing method of the broad ribbon, and a single-roll quenching method is adopted, wherein the width of a nozzle slot is 0.4-0.7 mm, the transversal width deviation of the nozzle slot is smaller than +/&#x2212;0.05 mm, the transversal flatness deviation of a cooling roll () is smaller than 0.02 mm, and the surface roughness Ra is smaller than 0.0005 mm.

Iron-based amorphous alloy broad ribbon and its manufacturing method

Growing community of inventors

Beijing, China

Pei Zhao