Implemented methods

The following methods are currently implemented in probly.

Representation

Second-order Distributions

These methods represent (epistemic) uncertainty by a second-order distribution over distributions.

Bayesian Neural Networks

[BCKW15]

Dropout

[GG16b]

DropConnect

[MNM+19]

Deep Ensembles

[LPB17b]

Evidential Deep Learning

[SKK18] [ASSR20]

Credal Sets

These methods represent (epistemic) uncertainty by a convex set of distributions.

Credal Ensembling

[NZD25]

Quantification

Upper / lower entropy

[AbellanKM06]

Generalized Hartley

[AbellanM00]

Entropy-based

[DHernandezLobatoDoshiVelezU18]

Distance-based

[SBCHullermeier24]

Conformal Prediction

These methods represent uncertainty by a set of predictions.

Split Conformal Prediction

[AB21]

Calibration

These methods adjust the model’s probabilities to better reflect the true probabilities.

Focal Loss

[LGG+17]

Label Relaxation

[LHullermeier21]

Temperature Scaling

[GPSW17b]