- Year: 2019
The main contribution of this paper is to explain why the overwhelming majority of hexapods don’t employ the rectilinear gallop using simple mechanical arguments.
- Year: 2018
Topics: AI, Uncertainty, Information Theory, Statistical Mechanics
The main contribution of this paper is to show that two intrinsic reward functions, the Causal Path Entropy and Empowerment, are equivalent only in deterministic environments. In non-deterministic environments, it is shown that the Causal Path Entropy has fundamental weaknesses compared to Empowerment. Moreover, the author demonstrates that the difference between Causal Path Entropy and Empowerment can’t be increased without diminishing Empowerment.
One motivation for writing this paper is that it addresses a question embedded in the following tweet by Shakir Mohamed, a research scientist at DeepMind: