EvidentialClassificationDecomposition

class probly.method.evidential.classification.EvidentialClassificationDecomposition(distribution: SecondOrderDistributionLike)[source]

Bases: CachingDecomposition, EpistemicDecomposition, Generic

Decomposition based on Evidential Deep Learning for classification.

Implements the single uncertainty quantity of EDL [SKK18]. For a Dirichlet distribution Dir(alpha) with K classes and S = sum_c alpha_c, the only uncertainty measure formally proposed by the paper is the vacuity:

  • Epistemic uncertainty: K / S, the “uncertainty mass” or vacuity from Dempster-Shafer / subjective-logic theory (Eq. 1 of the paper). Sensoy et al. explicitly call this the epistemic uncertainty of the classification (Sec. 3 of the paper).

The paper does not propose any aleatoric measure or aleatoric/epistemic decomposition; the subjective-opinion framework only splits between belief mass b_c = (alpha_c - 1) / S per class and the single vacuity term. Consequently this decomposition has only an epistemic slot, and its canonical notion is epistemic uncertainty.

Parameters:

distribution – A second-order distribution (e.g. a Dirichlet) or a sample from one. Requires a vacuity implementation registered for the input type.

canonical_notion[source]

alias of EpistemicUncertainty

property components: list[type[Notion]]

The components of the decomposition.

distribution: SecondOrderDistributionLike
property epistemic: EU

The epistemic uncertainty of the decomposition.

get(k[, d]) → D[k] if k in D, else d.  d defaults to None.
get_canonical() → Notion[source]

Return the canonical notion of the decomposition.

get_notion(notion: type[N]) → N[source]

Return the component corresponding to the given notion.

items() → a set-like object providing a view on D's items
keys() → a set-like object providing a view on D's keys
values() → an object providing a view on D's values

Examples using probly.method.evidential.classification.EvidentialClassificationDecomposition

Evidential on MNIST

Evidential on MNIST