Reliable spatial uncertainty evaluation of object detection models is of...
Calibrated confidence estimates obtained from neural networks are crucia...
Modern neural networks have found to be miscalibrated in terms of confid...
Understanding decisions made by neural networks is key for the deploymen...
On April 13th, 2019, OpenAI Five became the first AI system to defeat th...
We demonstrate that models trained only in simulation can be used to sol...
The purpose of this technical report is two-fold. First of all, it intro...
Dealing with sparse rewards is one of the biggest challenges in Reinforc...
Imitation learning has been commonly applied to solve different tasks in...
Developing control policies in simulation is often more practical and sa...
OpenAI Gym is a toolkit for reinforcement learning research. It includes...