The problem in classical planning

Unfortunately, the problem of assigning each sequence to a strategy isn’t addressed by Ramírez and Geffner’s (2009) approach. To resolve this problem, Ramírez and Geffner (2010) proposed modifying their original approach using a probabilistic formulation of plan recognition that was still based on classical planning.

Usage of probabilities

With this approach, they tried to find the probability of a goal given some observations, formulated as P(GO)P (G|O), where PP is the probability, GG is the goal, and OO is a set of observations.

Get hands-on with 1400+ tech skills courses.