IDiyas

For the Inventor, By the Inventor

New York, NY, United States of America

Ruoming Pang

USPTO Granted Patents = 39 

 

Average Co-Inventor Count = 6.3

ph-index = 4

Forward Citations = 48(Granted Patents)


Location History:

  • Jersey City, NJ (US) (2011 - 2012)
  • New York, NY (US) (2015 - 2024)

Company Filing History:

goldMedal39 out of 32,429 
 
Google Inc.
 patents

Years Active: 2011-2025

where 'Filed Patents' based on already Granted Patents

39 patents (USPTO):
1. 12482455 - Systems and methods for training dual-mode machine-learned speech recognition models (2025-11-25),
2. 12444408 - Two-pass end to end speech recognition (2025-10-14),
3. 12373666 - Convolution-augmented transformer models (2025-07-29),
4. 12361927 - Emitting word timings with end-to-end models (2025-07-15),
5. 12293276 - Neural architecture search with factorized hierarchical search space (2025-05-06),
6. 12190869 - Optimizing inference performance for conformer (2025-01-07),
7. 12183322 - Language agnostic multilingual end-to-end streaming on-device ASR system (2024-12-31),
8. 12175963 - Synthesis of speech from text in a voice of a target speaker using neural networks (2024-12-24),
9. 12154581 - Cascaded encoders for simplified streaming and non-streaming ASR (2024-11-26),
10. 12148444 - Synthesizing speech from text using neural networks (2024-11-19),
11. 12131244 - Hardware-optimized neural architecture search (2024-10-29),
12. 12118988 - Transducer-based streaming deliberation for cascaded encoders (2024-10-15),
13. 12112198 - Asynchronous distributed data flow for machine learning workloads (2024-10-08),
14. 12094453 - Fast emit low-latency streaming ASR with sequence-level emission regularization utilizing forward and backward probabilities between nodes of an alignment lattice (2024-09-17),
15. 12079703 - Convolution-augmented transformer models (2024-09-03),

Title: Innovations and Contributions of Ruoming Pang in Automatic Speech Recognition

Introduction

Ruoming Pang is an influential figure in the field of automatic speech recognition (ASR), based in New York, NY. With a remarkable number of 34 patents to his name, his work primarily revolves around optimizing the performance of ASR systems to enhance user experience across various applications.

Latest Patents

Among Ruoming Pang's latest innovations are two significant patents that focus on improving inference performance for ASR models. The first patent details a computer-implemented method involving a sequence of acoustic frames as input to an ASR model featuring a causal encoder and a decoder. This method generates a first higher-order feature representation for the corresponding acoustic frames and utilizes a Recurrent Neural Network (RNN) Attention-Performer module for linear attention.

The second patent introduces a language-agnostic multilingual end-to-end streaming on-device ASR system. This method also processes acoustic frames to generate higher-order feature representations and hidden representations. It classifies acoustic frames by identifying speech types and predicts the end of utterance tokens, showcasing a significant advancement in multilingual ASR systems.

Career Highlights

Ruoming Pang is currently affiliated with Google Inc., where he continues to push the boundaries of ASR technology. His contributions have positioned him as a pivotal innovator in this field, while his extensive patent portfolio reflects his deep commitment to research and development in speech recognition technologies.

Collaborations

Throughout his career, Ruoming Pang has collaborated with esteemed colleagues, including Tara N. Sainath and Yonghui Wu. These collaborations have facilitated the exchange of ideas and fostered innovative approaches to solving complex technical challenges in the realm of automated speech recognition.

Conclusion

Ruoming Pang's impactful work and numerous patents are instrumental in advancing automatic speech recognition systems. His continued dedication to innovation is vital for the future developments in the field, and his contributions are likely to have lasting effects on how we interact with technology through speech.

This text is generated by artificial intelligence and may not be accurate.
Please report any incorrect information to support@idiyas.com
Loading…