Experience · Software component

Explanation Method Selector

Software componentExperienceExperience & Human Oversightarc:ExplanationMethodSelector

A component that selects which explainability method(s) and visualizations to apply for a decision, based on user mode (novice or expert), investigation goal and decision context.

Responsibility. Selects the explainability methods applied to a decision for a given user and context.

Also known as: Multi-method explanation interface, Novice/expert mode switch

Direct neighbourhood (hover for relationship types)

Relationships

is invoked by dependency

Explanation Presenter Ch10.5

routes to dynamic

Design guidance

SHOULD offer multiple explanation methods rather than one universal method, since regulators, end users, developers and operators need different explanation types.
Novice mode SHOULD present a single default method selected by decision context; expert mode MAY expose all methods and visualization parameters.
Divergence between methods SHOULD be treated as a signal of complexity warranting further investigation or human oversight.

Classification

Patterns: Multi-method triangulation (SHAP, LIME, Anchors, DiCE)Novice mode single default methodExpert mode full method and visualization control
Technologies: IXAII
Quality attributes: Explainability (NIST AI RMF: explainable and interpretable)Flexibility (ISO/IEC 25010)Interaction capability (ISO/IEC 25010)
Risks mitigated: Universality assumption (single method for all users)

Sources

Ch10.5: T. Nguyen, "Human-over-the-Loop," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 10.5. ISBN: 9798244538229.