Tokyo, Japan

Byungha Chun

This inventor holds 20 USPTO granted patents and 15 published patent applications and 1 EPO patent, primarily in Text-To-Speech Technology. Top assignee: Google Inc.. Active years: 2013-2026.

IDiyas Innovation Intelligence. (2026). Inventor Profile: Byungha Chun. Retrieved from https://idiyas.com/inventor/byungha-chun

Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile

Calculating rank
USPTO Granted Patents = 20 

% Patents Active = 80.0

 

Average Co-Inventor Count = 5.1

ph-index = 3

Forward Citations = 384(Granted Patents)


Location History:

  • Surrey, GB (2013)
  • Epsom, GB (2015)
  • Epson, GB (2018)
  • Tokyo, JP (2022 - 2024)

Company Filing History:


Years Active: 2013-2026

Loading Chart...
Loading Chart...
Areas of Expertise:
Text-To-Speech
Speech Synthesis
Neural Networks
Multilingual Synthesis
Prosody Generation
Voice Cloning
Data Density Gradient
Few-Shot Adaptation
Duration Prediction
Tacotron
Speech Waveform Generation
Synthetic Training Data
20 patents (USPTO):Explore Patents

Title: The Innovations of Byungha Chun: Pioneering Speech Technology

Introduction

Byungha Chun, a prominent inventor based in Tokyo, Japan, has made significant contributions to the field of speech technology. With an impressive portfolio of 17 patents, Chun's work focuses on developing advanced systems that enhance speech synthesis and generation, paving the way for more expressive and accurate vocal outputs in various applications.

Latest Patents

Among his latest innovations, Chun has been instrumental in the creation of several groundbreaking patents. One such patent, titled "End-to-end speech waveform generation through data density gradient estimation," outlines methods and systems, including computer programs encoded on a storage medium, for generating waveforms conditioned on phoneme sequences. The method involves obtaining a phoneme sequence, processing it using an encoder neural network to derive a hidden representation, and ultimately generating a final waveform output. This approach allows for the dynamic updating of waveform outputs at various noise levels, ensuring the generated speech accurately reflects the intended phoneme sequence.

Another notable patent is the "Two-level speech prosody transfer," which enhances text-to-speech synthesis by capturing the intended prosody and target voice of input text utterances. This method leverages two text-to-speech models to create an intermediate synthesized speech representation that possesses the desired prosody before generating an output audio signal that reflects both the prosody and speaker characteristics of the target voice.

Career Highlights

Byungha Chun's career is marked by his tenure at Google Inc., where he has contributed to the advancement of speech technology. His research and innovations have earned him recognition in the field, and his patents highlight his commitment to enhancing human-computer interaction through improved speech recognition and synthesis methodologies.

Collaborations

Throughout his career, Chun has collaborated with talented colleagues, including Yu Zhang and Ye Jia. These collaborations have been essential in refining his innovative ideas and developing effective solutions that address complex challenges within the industry.

Conclusion

Byungha Chun continues to be a leading figure in speech technology innovation. His prolific patent portfolio and groundbreaking research contribute to the ongoing evolution of expressive speech synthesis. As advancements in this field progress, Chun's work stands as a testament to the impact that visionary inventors can have on enhancing communication technologies.

Profile summary based on public USPTO records.
Data Sources: USPTO Patent Grant XML, Patent Center, EPO & CIPO • Normalized by IDiyas Innovation Graph. Methodology & provenance architecturePlease report any incorrect information to [email protected]
Loading…