hmartiros

5 ms·

1. 1.▲
   by hmartiros 8y ago

   Skydio | Software Engineer - Skills SDK | Redwood City, CA | ONSITE VISA | Full-time Skydio makes the most advanced autonomous flying robots in the world today, and have an unprecedented opportunity to put autonomous UAVs to work at scale.
2. 2.▲
   by hmartiros 8y ago

   They aren't common in deep learning, but if you look to estimation problems like odometry, optimal control, and calibration, the typical approach is to build a least squares estimator that optimizes with a gauss-newton approximation to
3. 3.▲
   by hmartiros 8y ago

   This is really neat, thanks for sharing.
4. 4.▲
   by hmartiros 8y ago

   In my experience if you have even a little smoothness in your problem's cost manifold, taking advantage of gradients is invaluable to sample efficiency. Many losses which don't seem differentiable can be reformulated as such - you