ExplainabilityTransparencyTrustworthinesscs.AI
Towards Explainable AI Planning as a Service
Michael Cashmore, Anna Collins, Benjamin Krarup, Senka Krivic, Daniele Magazzeni, David Smith
August 14, 2019
arXiv.org
57 citations
4 influential
Key Takeaways
-
Explainable Planning can be added as a layer on top of existing systems: Rather than rebuilding planning systems from scratch, explainability can be designed as a "wrapper" around current planners, making it practical to add explanation capabilities to systems already in use without major overhauls.
-
Contrastive questions are a key way to explain AI decisions: The paper demonstrates that planners can effectively answer "why this plan and not that plan?" type questions, which helps users understand the reasoning behind AI-generated plans in a more intuitive way than technical descriptions.
-
This service-based approach has practical advantages and real limitations: While the framework makes explainability more accessible and easier to implement, the authors acknowledge open questions remain about how well this approach scales and what types of explanations it can realistically provide across different planning scenarios.
-
The research identifies multiple directions for future work: The paper highlights that Explainable Planning as a service is still emerging, with several unresolved questions about best practices, suggesting this is an active area where responsible AI implementation is still being developed.
Abstract
Explainable AI is an important area of research within which Explainable Planning is an emerging topic. In this paper, we argue that Explainable Planning can be designed as a service -- that is, as a wrapper around an existing planning system that utilises the existing planner to assist in answering contrastive questions. We introduce a prototype framework to facilitate this, along with some examples of how a planner can be used to address certain types of contrastive questions. We discuss the main advantages and limitations of such an approach and we identify open questions for Explainable Planning as a service that identify several possible research directions.