Abstraction for Deep Reinforcement Learning. (arXiv:2202.05839v1 [cs.LG])

We characterise the problem of abstraction in the context of deep
reinforcement learning. Various well established approaches to analogical
reasoning and associative memory might be brought to bear on this issue, but
they present difficulties because of the need for end-to-end differentiability.
We review developments in AI and machine learning that could facilitate their
adoption.

Source: https://arxiv.org/abs/2202.05839

webmaster

Related post