Jiachen Lian

Title: Innovations by Inventor Jiachen Lian in Speech Representation Learning

Introduction

Jiachen Lian is a prominent inventor based in Palo Alto, California, with a focus on advancing technologies in speech representation and synthesis. With an impressive portfolio of two patents, he has contributed significantly to the field of voice conversion and text-to-speech systems.

Latest Patents

Jiachen Lian's innovations are encapsulated in two key patents. The first patent, "Techniques for disentangled variational speech representation learning for zero-shot voice conversion," describes a method that enhances voice conversion by disentangling speaker identity and content. This method utilizes a shared encoder to generate embeddings, which are then further encoded and decoded to produce reconstructed speech outputs.

The second patent is titled "Text to speech synthesis without using parallel text-audio data." This unsupervised system innovatively maps input text to a phoneme sequence using a lexicon, which is then aligned to create a more accurate representation of speech. The Conditional Disentangled Sequential Variational Auto-encoder (C-DSVAE) refines the mel spectrogram output, facilitating sophisticated text-to-speech synthesis without the need for traditional parallel datasets.

Career Highlights

Jiachen Lian works as an engineer at Tencent America LLC, where he leverages his expertise to drive innovative research and development in voice technologies. His contributions have positioned him as a key player in the industry, particularly in the application of artificial intelligence to enhance human communication through speech.

Collaborations

While at Tencent America LLC, Jiachen has collaborated with notable colleagues including Chunlei Zhang and Dong Yu. These partnerships reflect a dynamic working environment focused on pushing the boundaries of what is possible in speech and audio technologies.

Conclusion

Jiachen Lian's innovative contributions to speech representation learning and text-to-speech synthesis are shaping the future of voice technology. His patents not only highlight his technical skills but also his commitment to enhancing how machines understand and generate human speech, paving the way for more intuitive and effective communication tools.