In simulation, we can replay trajectories and vary motions without needing to demonstrate them repeatedly with a real robot, enhancing efficiency in training.

We designed our technology to be performer-friendly and scalable, enabling content creation without the need for large motion capture stages.

In robotics, we are particularly focused on manipulation, which is currently one of the most unresolved challenges in the field.

We need to address the challenges of the Physical Turing Test, as current robotics struggle to perform tasks that humans can do effortlessly.

The performance capture technology we used for Alita involved capturing Rosa's body movements and facial expressions with multiple cameras to ensure we captured her emoting accurately.

Machine learning for control will enable robots to operate safely in messy environments and interact with humans effectively.

Data collection for robotics is challenging; we can't scrape the necessary data from the internet, and we rely on expensive teleoperation methods to teach robots.

Robotics is progressing rapidly, and with the integration of AI, we can expect robots to communicate and interact with humans more effectively.

I collaborated with a motion capture company to develop algorithms for camera calibration, leveraging the unique properties of geometric algebra to enhance the process.