Beijing, China

Ke Wei Sun

This inventor holds 31 USPTO granted patents and 11 published patent applications and 1 EPO patent. Top assignees: International Business Machines Corporation, Beijing Baidu Netcom Science and Technology Co., Ltd., Beijing Volcano Engine Technology Co., Ltd.. Active years: 2012-2025.

USPTO Granted Patents = 31 

% Patents Active = 80.6

 

Average Co-Inventor Count = 4.3

ph-index = 2

Forward Citations = 47(Granted Patents)


Company Filing History:


Years Active: 2012-2025

Loading Chart...
Loading Chart...
Areas of Expertise:
Metadata Server
Lakehouse Architecture
Artificial Intelligence
Deep Learning
Tensegrity Structures
Object Detection
Natural Language Processing
Cryptocurrency Transactions
Realtime Video Monitoring
Software Porting
Data Communication
Dual-Energy CT Scanning
31 patents (USPTO):Explore Patents

Title: Innovations and Achievements of Ke Wei Sun

Introduction

Ke Wei Sun is a prominent inventor located in Beijing, China, known for his significant contributions to the field of deep learning and neural network architecture. With a total of 26 patents to his name, Sun has been at the forefront of innovative technological advancements, steering critical developments in machine learning applications.

Latest Patents

One of his latest innovations is titled "Merge operations for darts," which presents a groundbreaking one-shot neural architecture search method known as MergeNAS. This inventive approach merges various types of convolutions into a single operation, effectively reducing the search cost to approximately half a GPU-day. Additionally, it addresses the overfitting challenges commonly associated with the traditional Differentiable Architecture Search (DARTS) approach by minimizing redundant parameters.

Another notable patent is "Search space exploration for deep learning." This invention includes systems and methods to extract meta features of a dataset crucial for training in deep learning applications. The method outlines a process for selecting an initial search space that defines a type of deep learning architecture representation that specifies hyperparameters for multiple neural network architectures. By employing a search strategy, one neural network architecture is selected based on evaluation results, leading to the generation of a new search space with updated hyperparameters using an evolutionary algorithm.

Career Highlights

Throughout his career, Ke Wei Sun has made significant impacts while working at leading companies in the technology sector. Notable organizations include IBM and Beijing Baidu Netcom Science and Technology Co., Ltd. His work in these prestigious companies has enabled him to develop and refine his innovative ideas, pushing boundaries in the realm of neural networks.

Collaborations

Within his professional journey, Sun has collaborated with esteemed colleagues such as Shiqi Zhao and Tingting Li. These collaborations have enriched his research and innovation efforts, fostering an environment for creativity and technological advancements.

Conclusion

Ke Wei Sun continues to be a visionary inventor in the realms of deep learning and artificial intelligence. His ongoing commitment to innovation is demonstrated by his extensive patent portfolio and collaborative efforts within top technology companies. As the landscape of machine learning evolves, Sun’s contributions remain pivotal in shaping the future of technology.

This text is generated by artificial intelligence and may not be accurate.
Please report any incorrect information to [email protected]
Loading…