How simulator exploitation, a dual of over-optimization, in AI is the canary in the coal mine for what negative implications could come from weakly-bounded, data-driven iterative systems (RL).
I've spent a while using RL for walking robots and now I predict RL will never be used on real robots for any control task. All these RL implementations on Mujoco are a waste of time, conferences should have higher standards for accepting papers
I've spent a while using RL for walking robots and now I predict RL will never be used on real robots for any control task. All these RL implementations on Mujoco are a waste of time, conferences should have higher standards for accepting papers