Jianan Wang

Title: Jianan Wang and His Patent on Gated Linear Contextual Bandits.

Introduction

Jianan Wang is an inventor associated with Hong Kong, China, and Gdm Holding LLC. His listed patent work concerns training neural networks to control real-world agents.

Latest Patents

Wang is credited with 1 patent: “Gated Linear Contextual Bandits.” The patent describes methods, systems, and computer-readable media for training a neural network to perform a task by controlling an agent in a real-world environment. The approach first trains the network using a task-specific objective in a simulated environment. It then uses data from the real agent’s interactions with its environment to refine the network, jointly optimizing a self-supervised objective and a second task-specific objective.

Career Highlights

The available information identifies Wang’s work with Gdm Holding LLC and highlights his contribution to machine-learning methods for real-world agent control. The patent combines simulation-based training with learning from real-world data.

Collaborations

Wang’s listed coworkers include Eren Sezener and Joel William Veness. Their names are associated with the same patent information.

Conclusion

Jianan Wang’s listed invention focuses on combining simulated training and real-world learning to improve neural-network control of agents performing specific tasks.