We give a brief presentation of the maximization of path entropy (Maximum Caliber) formalism as an approach to inference over trajectories. We develop its continuoustime version involving path integrals and present an identity between expectations over the most unbiased probability distribution of trajectories. This relation between expectations can be used to solve for the Lagrange multipliers in a Maximum Caliber problem without the need for the dynamical partition function.
1 May 2016
Volume: 720 Issue: 1 Pages: 012006
Journal of Physics: Conference Series