Marcus Hutter

Title: Marcus Hutter and His Patent in Machine Learning

Introduction

Marcus Hutter is an inventor associated with London, Great Britain. His listed patent work concerns machine-learning methods for training neural networks to control agents operating in real-world environments. The supplied information lists 1 patent and identifies Gdm Holding LLC as his company.

Latest Patents

The listed patent is “Gated linear contextual bandits.” It describes methods, systems, and computer programs for training a neural network to control a real-world agent performing a particular task.

The approach first trains the network using a task-specific objective based on the agent’s performance in a simulated environment. It then uses data from the agent’s real-world interactions to refine the network’s parameters by jointly optimizing a self-supervised objective and a second task-specific objective. The self-supervised component evaluates internal representations on a task performed using real-world data.

Career Highlights

The patent connects simulated training with learning from real-world data, combining task-specific optimization with self-supervised learning. Gdm Holding LLC is identified as Hutter’s company in the information provided.

Collaborations

The supplied information names Eren Sezener and Joel William Veness as Hutter’s coworkers. Their inclusion indicates professional associations, but no specific roles or contributions to the patent are provided.

Conclusion

Marcus Hutter’s listed patent focuses on adapting neural-network training from simulated settings to real-world interactions, with self-supervised and task-specific objectives working together.