quantification¶
Quantification methods for uncertainty.
Submodules¶
Uncertainty decomposition methods. |
|
Measures the uncertainty of a given representation. |
|
Definition of different notions of uncertainty. |
Classes¶
Protocol for uncertainty quantifications. |
|
Protocol for uncertainty quantification methods. |
Functions¶
Compute the conditional entropy of a distribution. |
|
Compute the mean spectral entropy within groups of embedding samples [WGD+26]. |
|
Measure conformal-set uncertainty by set size. |
|
Decomposes the uncertainty of a given representation. |
|
Compute the Dempster-Shafer uncertainty score of a distribution over logits. |
|
Compute the entropy of a distribution. |
|
Compute the entropy of the expected value of a second-order distribution. |
|
Compute the expected value of one minus the max probability under a second-order distribution. |
|
Compute the expected gap between each sample's max probability and its probability on the BMA argmax. |
|
Compute one minus the max probability of the expected value of a second-order distribution. |
|
Mean over members of \(\| h_k - s\, e_{\arg\max_j h_{k,j}} \|_2^2\); class axis is last. |
|
Measures the uncertainty of a given representation. |
|
Measures the uncertainty of a given representation without attempting to decompose it. |
|
Compute the distance-based epistemic uncertainty of a second-order distribution. |
|
Compute the mutual information of a distribution. |
|
A simple switch-case mechanism based on a dictionary. |
|
Quantifies the uncertainty of a given representation. |
|
Measure uncertainty for samples via their sample variance. |
|
Compute spectral uncertainty as Von Neumann entropy of an RBF kernel [WGD+26]. |
|
Measure uncertainty for samples via the variance of their total logits (logits summed across members). |
|
Compute the vacuity of a second-order distribution. |