Active Inference for Stochastic Control

Active inference has emerged as an alternative approach to control problems given its intuitive (probabilistic) formalism. However, despite its theoretical utility, computational implementations have largely been restricted to low-dimensional, deterministic settings. This paper highlights that this is a consequence of the inability to adequately model stochastic transition dynamics, particularly when an extensive policy (i.e., action trajectory) space must be evaluated during planning. Fortunately, recent advancements propose a modified planning algorithm for finite temporal horizons. We build upon this work to assess the utility of active inference for a stochastic control setting. For this, we simulate the classic windy grid-world task with additional complexities, namely: 1) environment stochasticity; 2) learning of transition dynamics; and 3) partial observability. Our results demonstrate the advantage of using active inference, compared to reinforcement learning, in both deterministic and stochastic settings.

[1]  Mark Miller,et al.  Losing Ourselves: Active Inference, Depersonalization, and Meditation , 2020, Frontiers in Psychology.

[2]  Beren Millidge,et al.  On the Relationship Between Active Inference and Control as Inference , 2020, IWAI.

[3]  Karl J. Friston,et al.  Active inference on discrete state-spaces: A synthesis , 2020, Journal of mathematical psychology.

[4]  Gordon Cheng,et al.  Active inference body perception and action for humanoid robots , 2019, ArXiv.

[5]  Richard S. Sutton,et al.  Reinforcement Learning: An Introduction , 1998, IEEE Trans. Neural Networks.

[6]  Karl J. Friston,et al.  Active Inference: Demystified and Compared , 2019, Neural Computation.

[7]  Karl J. Friston,et al.  Reinforcement Learning or Active Inference? , 2009, PloS one.

[8]  Karl J. Friston,et al.  Deep active inference agents using Monte-Carlo methods , 2020, NeurIPS.

[9]  Tim Verbelen,et al.  Bayesian policy selection using active inference , 2019, ICLR 2019.

[10]  Karl J. Friston,et al.  The relationship between dynamic programming and active inference: the discrete, finite-horizon case , 2020, ArXiv.

[11]  Karl J. Friston,et al.  Future climates: Markov blankets and active inference in the biosphere , 2020, Journal of the Royal Society Interface.

[12]  Karl J. Friston,et al.  Planning and navigation as active inference , 2017, Biological Cybernetics.

[13]  Otto van der Himst,et al.  Deep Active Inference for Partially Observable MDPs , 2020, IWAI.

[14]  Karl J. Friston,et al.  Sophisticated Inference , 2020, Neural Computation.

[15]  Karl J. Friston,et al.  Active Inference: A Process Theory , 2017, Neural Computation.

[16]  Karl J. Friston The free-energy principle: a unified brain theory? , 2010, Nature Reviews Neuroscience.

[17]  Karl J. Friston,et al.  Active inference and agency: optimal control without cost functions , 2012, Biological Cybernetics.

[18]  Karl J. Friston,et al.  Morphogenesis as Bayesian inference: A variational approach to pattern formation and control in complex biological systems. , 2020, Physics of life reviews.