Release 1.3.6 - dependency fixes

This release mainly consists of fixes in third-party dependencies and minor bugfixes within the library.

Full Changelog: r1.3.5...r1.3.6

Release 1.3.5 for feature request #40

Implements feature request #40 by merging pull request #41 : add new "figure" parameter to diagram drawing routines to allow for consumption of pre-existing matplotlib figure objects.

Release 1.3.4 - bugfix for issue #38

Resolve issue #38 by pull request #39 .

Release 1.3.3 - bugfix for issue #35

Resolve issue #35

Release 1.3.2 - bugfix for issue #31

Resolve issue #31

Release 1.3.1 - minor path updates and ownership transfer

Ownership of this repository has changed to the "EFS-OpenSource" team.
All paths for repository checkout and API documentation have been changed.
This release updates the relevant information on PyPI.

Release 1.3.0 - new methods for regression uncertainty calibration

Within this release, we provide a new package netcal.regression to enable recalibration of probabilistic regression tasks.

• Regression calibration methods: train and infer methods to rescale the uncertainty of probabilistic regression models

• New package: netcal.regression with regression calibration methods:

  • Isotonic Regression (netcal.regression.IsotonicRegression)
  • Variance Scaling (netcal.regression.VarianceScaling)
  • GP-Beta (netcal.regression.GPBeta)
  • GP-Normal (netcal.regression.GPNormal)
  • GP-Cauchy (netcal.regression.GPCauchy)

• Implement netcal.regression.GPNormal method with correlation estimation and recalibration

• Restructured netcal.metrics package to distinguish between (semantic) confidence calibration in netcal.confidence and regression uncertainty calibration in netcal.regression:

  • Expected Calibration Error (ECE - netcal.confidence.ECE)
  • Maximum Calibration Error (MCE - netcal.confidence.MCE)
  • Average Calibration Error (ACE - netcal.confidence.ACE)
  • Maximum Mean Calibration Error (MMCE - netcal.confidence.MMCE)
  • Negative Log Likelihood (NLL - netcal.regression.NLL)
  • Prediction Interval Coverage Probability (PICP - netcal.regression.PICP)
  • Pinball loss (netcal.regression.PinballLoss)
  • Uncertainty Calibration Error (UCE - netcal.regression.UCE)
  • Expected Normalized Calibration Error (ENCE - netcal.regression.ENCE)
  • Quantile Calibration Error (QCE - netcal.regression.QCE)

• Added new types of reliability diagrams to visualize regression calibration properties:

  • Reliability Regression diagram to visualize calibration for different quantile levels (preferred - netcal.presentation.ReliabilityRegression)
  • Reliability QCE diagram to visualize QCE over stddev (netcal.presentation.QCE)

• Updated examples

• Minor bugfixes

• Use library tikzplotlib within the netcal.presentation package to enable a direct conversion of matplotlib.Figure objects to Tikz-Code (e.g., can be used for LaTeX figures)

Release 1.2.1 - binary classification bugfix

Fixed #17 where correct identification of binary classification case has failed in reliability diagram/metrics.

Release 1.2.0 - major updates for scaling methods

Update repository by content described in our IV 2021 paper: "Bayesian Confidence Calibration for Epistemic Uncertainty Modelling"

• Bayesian confidence calibration: train and infer scaling methods using variational inference (VI) and MCMC sampling
• New metrics: MMCE and PICP
• New regularization methods: MMCE and DCA
• Updated examples
• Switched license from MPL2 to APL2

Release 1.1.3 - minor bugfixes

• update Numpy version number requirement (fixed issue #6)
• removed logger as member variable from base class to prevent pickling issues (fixed issue #9)
• updated reference to our CVPR 2020 publication