1. What are the contributions mentioned in the paper "Interval change-point detection for runtime probabilistic model checking" ?
To address this limitation, the authors introduce an efficient interval change-point detection method, and they integrate it with a state-of-the-art Bayesian estimator with imprecise priors.. Their experimental results show that the resulting end-to-end Bayesian approach to change-point detection and estimation of interval Markov chain parameters handles effectively a wide range of sudden changes in parameter values, and supports runtime probabilistic model checking under parametric uncertainty.
read more
2. What have the authors stated for future works in "Interval change-point detection for runtime probabilistic model checking" ?
As future work, the authors plan to extend iCPD with support for other patterns of change ( e. g., waves, triangles ), and to investigate principled mechanisms of eliciting the IPSP priors.
read more
![Figure 6: iCPD overheads in scenarios A and B showing computation time (left), change point estimation error (middle), and MCMC diagnostics acceptance rate [23] (right) over the number of conducted MCMC trials.](/figures/figure-6-icpd-overheads-in-scenarios-a-and-b-showing-246faplw.webp)
![Table 1: Configurations and iCPD results over nine iDTMC scenarios. In all scenarios, the IPSP prior parameters are set to [𝑝𝑖 𝑗 (0) , 𝑝𝑖 𝑗 (0) ] = [0.2, 0.4] and [𝑛 (0) , 𝑛 (0) ] = [10, 300]. The scenario ID is associated with the corresponding subfigure in Fig 3 (Notations ś 𝑥 : change point; 𝑥𝑡 : iCPD trigger point; 𝑥 : iCPD estimated change point;𝑘/𝑙 : ratio of accepted samples over the total number of MCMC trials; ⟨𝑥1, . . . , 𝑥𝑘 ⟩: MCMC sample sequence of the change point; ⟨𝑏1, . . . , 𝑏𝑘 ⟩: sample sequence of transition probability 𝑏; N/A: iCPD not triggered).](/figures/table-1-configurations-and-icpd-results-over-nine-idtmc-22zka73t.webp)
