Singapore, Singapore

Guangsen Wang

USPTO Granted Patents = 2 

Average Co-Inventor Count = 4.0

ph-index = 1

Forward Citations = 2(Granted Patents)


Company Filing History:


Years Active: 2023-2024

Loading Chart...
2 patents (USPTO):Explore Patents

Title: Innovations of Guangsen Wang in Speech Recognition and Dialog Systems

Introduction

Guangsen Wang is a notable inventor based in Singapore, recognized for his contributions to the fields of speech recognition and dialog systems. With a total of two patents to his name, he has developed innovative frameworks that enhance the functionality and efficiency of task-oriented systems.

Latest Patents

One of Guangsen Wang's latest patents is titled "Systems and methods for an end-to-end evaluation and testing framework for task-oriented dialog systems." This invention provides a software framework designed for evaluating and troubleshooting real-world task-oriented bot systems. The evaluation framework includes a generator that infers dialog acts and entities from bot definitions, generating test cases for the system through model-based paraphrasing. Additionally, it features a simulator for task-oriented dialog user simulation, supporting both regression testing and end-to-end evaluation. The framework also incorporates a remediator to analyze and visualize simulation results, remedy identified issues, and offer actionable suggestions for improving the task-oriented dialog system.

Another significant patent is "Systems and methods for a multilingual speech recognition framework." This invention introduces an Adapt-and-Adjust (A2) mechanism for multilingual speech recognition models, combining adaptation and adjustment methods into an integrated end-to-end training process. This approach aims to improve the models' generalization and mitigate long-tailed issues. The framework utilizes a multilingual language model, mBERT, which is converted into an autoregressive transformer decoder. A cross-attention module is added to the encoder on top of mBERT's self-attention layer to explore the acoustic space alongside the text space. The joint training of the encoder and mBERT decoder effectively bridges the semantic gap between speech and text.

Career Highlights

Guangsen Wang has worked with prominent companies, including Salesforce.com, Inc. and Salesforce, Inc. His experience in these organizations has contributed to his expertise in developing advanced technologies in speech recognition and dialog systems.

Collaborations

Some of his notable coworkers include Chu Hong Hoi and Genta Indra Winata, who have collaborated with him on various projects.

Conclusion

Guangsen Wang's innovative patents and career achievements highlight his significant contributions to the fields of speech recognition and dialog systems. His work continues to influence advancements in technology, making him a key figure in these domains.

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