Applications Technology (apptek), LLC

Apptek, Inc.

Ebay, Incorporated

Mudar Yaghi

Jintao Jiang

Hassan Sawaf

Evgeny Matusov

Mohammad Shihadah

Darius Ferdows

Jinato Jiang

Nick Rossenbach

Yasar Yaghi

A system for translating speech from at least two source languages into another target language provides direct speech to target language translation. The target text is converted to speech in the target language through a TTS system. The system simplifies speech recognition and translation process by providing direct translation, includes mechanisms described herein that facilitate mixed language source speech translation, and punctuating output text streams in the target language. It also in some embodiments allows translation of speech into the target language to reflect the voice of the speaker of the source speech based on characteristics of the source language speech and speaker's voice and to produce subtitled data in the target language corresponding to the source speech. The system uses models having been trained using (i) encoder-decoder architectures with attention mechanisms and training data using TTS and (ii) parallel text training data in more than two different languages.

System and method for direct speech translation system

A system and method enable one to set a target duration of a desired synthesized utterance without removing or adding spoken content. Without changing the spoken text, the voice characteristics may be kept the same or substantially the same. Silence adjustment and interpolation may be used to alter the duration while preserving speech characteristics. Speech may be translated prior to a vocoder step, pursuant to which the translated speech is constrained by the original audio duration, while mimicking the speech characteristics of the original speech.

Method and apparatus for forced duration in neural speech synthesis

A mobile device is programmed with an application that uses the mobile device's camera, accelerometer and microphone to enable a parent, coach or player to use it as a tool to diagnose a concussion. The tool may diagnose concussion on the basis of one or multiple factors that are scored, for example the player's balance, eye movement, speech responses to questions, button pressing response time, and other information about the location of the impact. A mobile device may be equipped with speech recognition and voice prompting to enable a concussion examination of a player to be administered by another player or coach to the injured player without significant effort by the injured player or helper. Each test may be scored, by itself or against one or more baselines for the injured player to develop an overall score and likelihood of a concussion. When the coach thinks there is a concussion, he/she can use the application to help find a doctor. When the coach is not sure about the test results, he/she can send the machine learning annotated data to an online doctor for reviewing.

Systems and methods for diagnosing and analyzing concussions

Deep Neural Networks (DNN) are time shifted relative to one another and trained. The time-shifted networks may then be combined to improve recognition accuracy. The approach is based on an automatic speech recognition (ASR) system using DNN and using time shifted features. Initially, a regular ASR model is trained to produce a first trained DNN. Then a top layer (e.g., SoftMax layer) and the last hidden layer (e.g., Sigmoid) are fine-tuned with same data set but with a feature window left- and right-shifted to create respective second and third left-shifted and right-shifted DNNs. From these three DNN networks, four combination networks may be generated: left- and right-shifted, left-shifted and centered, centered and right-shifted, and left-shifted, centered, and right-shifted. The centered networks are used to perform the initial (first-pass) ASR. Then the other six networks are used to perform rescoring. The resulting are combined using ROVER (recognizer output voting error reduction) or another technique to improve recognition performance as compared to the centered DNN by itself.

Centered, left- and right-shifted deep neural networks and their combinations

A hybrid frame, phone, diphone, morpheme, and word-level Deep Neural Networks (DNN) in model training and applications-is based on training a regular ASR system, which can be based on Gaussian Mixture Models (GMM) or DNN. All the training data (in the format of features) are aligned with the transcripts in terms of phonemes and words with the timing information and new features are formed in terms of phonemes, diphones, morphemes, and up to words. Regular ASR produces a result lattice with timing information for each word. A feature is then extracted and sent to the word-level DNN for scoring Phoneme features are sent to corresponding DNNs for training. Scores are combined to form the word level scores, a rescored lattice and a new recognition result.

Hybrid phoneme, diphone, morpheme, and word-level deep neural networks

Phoneme images are created for keywords and audio files. The keyword images and audio file images are used to identify keywords within the audio file when the phoneme images match. Confidence scores may be determined corresponding to the match. Audio around the keywords may be stored and processed with an automatic speech recognition (ASR) program to verify the keyword match and provide textual and audio context to where the keyword appears within speech.

Method and apparatus for keyword speech recognition

A system and method for hybrid machine translation approach is based on a statistical transfer approach using statistical and linguistic features. The system and method may be used to translate from one language into another. The system may include at least one database, a rule based translation module, a statistical translation module and a hybrid machine translation engine. The database(s) store source and target text and rule based language models and statistical language models. The rule based translation module translates source text based on the rule based language models. The statistical translation module translates source text based on the statistical language models. A hybrid machine translation engine, having a maximum entropy algorithm, is coupled to the rule based translation module and the statistical translation module and is capable of translating source text into target text based on the rule based and statistical language models.

Growing community of inventors

McLean, VA, United States of America

Mudar Yaghi