Related to: [Name of Request for Research]
In the paper multi-goal reinforcement learning challenging robotics environments and request for research
(https://arxiv.org/pdf/1802.09464.pdf) they are showing the demo of their proposed approach in physical robot.
However, there is no detail in the paper how they transfer the learning from virtual environment to actual physical robot. Is it designed and customized for a special brand of robot or it can be used for any robot!
There is a missing part in every paper that I read in this area which they dont reveal these kind of information. If we can train the model in virtual environment then how we can transfer the learning or deploy the model in an actual model.