A hybrid approach to dialogue management based on probabilistic rules

Pierre Lison
2015 Computer Speech and Language  
We present a new modelling framework for dialogue management based on the concept of probabilistic rules. Probabilistic rules are defined as structured mappings between logical conditions and probabilistic effects. They function as high-level templates for probabilistic graphical models and may include unknown parameters whose values are estimated from data using Bayesian inference. Thanks to their use of logical abstractions, probabilistic rules are able to encode the probability and utility
more » ... dels employed in dialogue management in a compact and human-readable form. As a consequence, they can reduce the amount of dialogue data required for parameter estimation and allow system designers to directly incorporate their expert domain knowledge into the dialogue models. Empirical results of a user evaluation in a human-robot interaction task with 37 participants show that a dialogue manager structured with probabilistic rules outperforms both purely hand-crafted and purely statistical methods on a range of subjective and objective quality metrics. The framework is implemented in a software toolkit called OpenDial, which can be used to develop various types of dialogue systems based on probabilistic rules. 2 The variables x are variables in the sense of first-order logic and should not be confused with the random variables of the probabilistic model. 3 In other words, the probability values must satisfy to 0 ≤ θ i,j ≤ 1 for all values i, j and m i j=1 θ i,j = 1 for every branch of a given rule.
doi:10.1016/j.csl.2015.01.001 fatcat:5j67rdwmvfdh5jkbsieoecqnou