loss: Loss and regularisation

Loss functions and regularisations.

The loss submodule is a collection of implementations of differentiable functions that map from arbitrary inputs to scalar-valued outputs. These scalar outputs can provide a starting point for a backward pass through a differentiable program model. Functionality is provided for various measures of interest to functional brain mapping and other contexts.

Helper wrappers allow packaging of multiple loss objectives in a single call. Each wrapped objective can be selectively applied to a subset of input tensors using LossApply, LossArgument, and LossScheme functionality.

• Loss: Base class for scalar-valued losses
  • Loss
• ParameterisedLoss: Extensible class for custom parameterised losses
  • ParameterisedLoss
• MSELoss: Mean squared error
  • MSELoss
• NormedLoss: Normed parameter regularisation
  • NormedLoss
• identity: Identity function
  • identity()
• zero: Zero function
  • zero()
• difference: Elementwise difference
  • difference()
• constraint_violation: Soft constraints
  • constraint_violation
    • constraint_violation()
  • ConstraintViolationLoss
    • ConstraintViolationLoss
• unilateral_loss: Unilateral penalties
  • unilateral_loss
    • unilateral_loss()
  • UnilateralLoss
    • UnilateralLoss
• hinge_loss: SVM hinge loss
  • hinge_loss
    • hinge_loss()
  • HingeLoss
    • HingeLoss
• smoothness: Backwards differences
  • smoothness
    • smoothness()
  • SmoothnessLoss
    • SmoothnessLoss
• bimodal_symmetric: Minimal distance from 2 modes
  • bimodal_symmetric
    • bimodal_symmetric()
  • BimodalSymmetricLoss
    • BimodalSymmetricLoss
• det_gram: Gramian determinant
  • det_gram
    • det_gram()
  • GramDeterminantLoss
    • GramDeterminantLoss
• log_det_gram: Gram log-determinant loss
  • log_det_gram
    • log_det_gram()
  • GramLogDeterminantLoss
    • GramLogDeterminantLoss
• entropy: Categorical entropy
  • entropy
    • entropy()
    • entropy_logit()
  • EntropyLoss
    • EntropyLoss
    • EntropyLogitLoss
• kl_divergence: Kullback-Leibler divergence
  • kl_divergence
    • kl_divergence()
    • kl_divergence_logit()
  • KLDivergenceLoss
    • KLDivergenceLoss
    • KLDivergenceLogitLoss
• js_divergence: Jensen-Shannon divergence
  • js_divergence
    • js_divergence()
    • js_divergence_logit()
  • JSDivergenceLoss
    • JSDivergenceLoss
    • JSDivergenceLogitLoss
• bregman_divergence: Bregman divergences
  • bregman_divergence
    • bregman_divergence()
    • bregman_divergence_logit()
  • BregmanDivergenceLoss
    • BregmanDivergenceLoss
    • BregmanDivergenceLogitLoss
• equilibrium: Equilibrium loss
  • equilibrium
    • equilibrium()
    • equilibrium_logit()
  • EquilibriumLoss
    • EquilibriumLoss
    • EquilibriumLogitLoss
• second_moment: Second moments
  • second_moment
    • second_moment()
    • second_moment_centred()
  • SecondMomentLoss
    • SecondMomentLoss
    • SecondMomentCentredLoss
• auto_tol: Significance tolerance
  • auto_tol()
• batch_corr: Batch-axis correlation
  • batch_corr
    • batch_corr()
  • BatchCorrelationLoss
    • BatchCorrelationLoss
• qcfc: QC-FC measures and loss
  • qcfc
    • qcfc()
  • QCFCLoss
    • QCFCLoss
• reference_tether: Spatial tether to reference points
  • reference_tether
    • reference_tether()
  • ReferenceTetherLoss
    • ReferenceTetherLoss
• interhemispheric_tether: Inter-hemispheric tethering loss
  • interhemispheric_tether
    • interhemispheric_tether()
  • InterhemisphericTetherLoss
    • InterhemisphericTetherLoss
• compactness: Compactness
  • compactness
    • compactness()
  • CompactnessLoss
    • CompactnessLoss
• dispersion: Vector dispersion
  • dispersion
    • dispersion()
  • DispersionLoss
    • DispersionLoss
• multivariate_kurtosis: Time series stationarity
  • multivariate_kurtosis
    • multivariate_kurtosis()
  • MultivariateKurtosis
    • MultivariateKurtosis
• connectopy: Generalised connectopy
  • connectopy
    • connectopy()
  • ConnectopyLoss
    • ConnectopyLoss
• modularity: Relaxed modularity
  • modularity
    • modularity()
  • ModularityLoss
    • ModularityLoss
• sum_scalarise: Sum scalarisation
  • sum_scalarise()
• mean_scalarise: Mean scalarisation
  • mean_scalarise()
• meansq_scalarise: Squared mean scalarisation
  • meansq_scalarise()
• max_scalarise: Maximum-value scalarisation
  • max_scalarise()
• norm_scalarise: Norm scalarisation
  • norm_scalarise()
• vnorm_scalarise: Vector norm scalarisation
  • vnorm_scalarise()
• wmean_scalarise: Weighted mean scalarisation
  • wmean_scalarise()
• selfwmean_scalarise: Self-weighted mean scalarisation
  • selfwmean_scalarise()
• LossApply: Selectively apply loss to parameters
  • LossApply
• LossScheme: Scheme for multiple losses
  • LossScheme