srly, at least the closest thing we have today.
The bottleneck to on-robot reinforcement learning is good, scalable reward prediction. Robometer is a massive step in that direction, and the authors have been wonderfully open as well, releasing a large dataset and continuing to improve their model post release.
Thanks to @aliangdw @yigitkkorkmaz and @Jesse_Y_Zhang for joining me and @DJiafei!
Natural language input is going to be the real bottleneck for adoption when Robots are deployed. There's no point in having a robot if you need someone with a PHD to operate it
The bottleneck to on-robot reinforcement learning is good, scalable reward prediction. Robometer is a massive step in that direction, and the authors have been wonderfully open as well, releasing a large dataset and continuing to improve their model post release.
Thanks to @aliangdw @yigitkkorkmaz and @Jesse_Y_Zhang for joining me and @DJiafei!
srly, at least the closest thing we have today.
The bottleneck to on-robot reinforcement learning is good, scalable reward prediction. Robometer is a massive step in that direction, and the authors have been wonderfully open as well, releasing a large dataset and continuing to improve their model post release.
Thanks to @aliangdw @yigitkkorkmaz and @Jesse_Y_Zhang for joining me and @DJiafei!