Man pages for mgoulet847/tagi
Tractable Approximate Gaussian Inference in Neural Networks
| activationFunIndex | Assign ID to activation functions |
| backwardHiddenStateUpdate | Backward hidden states update |
| backwardParameterUpdate | Backward parameters update |
| batchDerivative | One iteration of the TAGI with derivative calculations |
| BH | Price of 506 Boston houses. |
| buildCzp | Reformat covariance matrix between units and parameters |
| buildCzz | Reformat covariance matrix between units of the previous and... |
| catParameters | Concatenate parameters |
| compressParameters | Compress parameters |
| compressStates | Compress states |
| computeError | Compute error |
| covariance | Indices for covariances in the neural network |
| covarianceCzp | Covariance matrices between units and parameters |
| covarianceCzz | Covariance matrix between units of the previous and current... |
| covarianceSa | Calculate variance of activated units |
| covarianceSz | Covariance matrix of units |
| covdx | Covariance between derivatives and hidden states |
| createDevCellarray | States initialization (unit matrices) |
| createInitCellwithArray | Initialization (matrix of lists) |
| createStateCellarray | States initialization (zero-matrices) |
| denormalize | Denormalize data |
| derivative | Derivative calculation |
| extractParameters | Extract parameters |
| extractStates | Extract states |
| fcCombinaisonDnode | Combination of products of first derivative (iterations on... |
| fcCombinaisonDweight | Combination of products of first derivative (iterations on... |
| fcCombinaisonDweightNode | Combination of squared products of first derivative |
| fcCombinaisonDweightNodeAll | All possible combinations of products of first derivatives |
| fcCovaddddddw | Covariance between first and second derivatives from... |
| fcCovawaa | Covariance between activation units and weights |
| fcCovaz | Covariance between activation and hidden units |
| fcCovdaddd | Covariance between first and second derivatives from... |
| fcCovDlayer | Covariance between products of derivatives and weights |
| fcCovdwd | Covariance between derivatives and weights*derivatives |
| fcCovdwddd | Covariance between derivatives and weights |
| fcCovdz | Covariance between derivatives and hidden Units |
| fcCovwdo2wdiwdi | Covariance between products in (same) next and current layers |
| fcCovwdowdi2 | Covariance between next layer product and current layer... |
| fcCovwdowdiwdi | Covariance between next layer product and current layer... |
| fcCwdowdowdiwdi | Covariance between next layer multiplied products and current... |
| fcCwdowdowdiwdi_4hl | Covariance between next layer multiplied products and current... |
| fcCwdowdowwdi2 | Covariance between next layer multiplied products and current... |
| fcCwdowdowwdi2_3hl | Covariance between next layer multiplied products and current... |
| fcDerivative | Derivatives for fully connected layers |
| fcDerivative2 | Second derivatives for fully connected layers |
| fcDerivative3 | Products of first derivatives multiplied to second derivative... |
| fcDerivative4 | Products of first derivatives multiplied to products of first... |
| fcDerivative5 | Products of first derivatives multiplied to products of first... |
| fcHiddenStateBackwardPass | Backpropagation (states' deltas) for fully connected layers... |
| fcHiddenStateBackwardPassB1 | Backpropagation (states' deltas) for fully connected layers... |
| fcMeanDlayer2array | Mean of weights times derivatives products terms ((wdo*wdo) x... |
| fcMeanDlayer2row | Mean of weights times derivatives products terms squared (wdo... |
| fcMeanVar | Mean and covariance vectors of units (many observations) |
| fcMeanVarB1 | Mean and covariance vectors of units (one observation) |
| fcMeanVarDlayer | Mean and variance of weights times derivatives products terms |
| fcMeanVarDnode | Mean and covariance of derivatives |
| fcParameterBackwardPass | Backpropagation (parameters' deltas) for fully connected... |
| fcParameterBackwardPassB1 | Backpropagation (parameters' deltas) for fully connected... |
| feedBackward | Backpropagation |
| feedForward | Forward uncertainty propagation |
| feedForwardPass | Forward uncertainty propagation for derivative calculation |
| forwardHiddenStateUpdate | Last hidden layer states update |
| globalParameterUpdate | Backpropagation (parameters update) |
| hiddenStateBackwardPass | Backpropagation (states' deltas) |
| initialization | Network initialization |
| initialization_net | Network initialization |
| initializeInputs | Input initialization |
| initializeStates | States initialization |
| initializeWeightBias | Weights and biases initialization |
| initializeWeightBiasD | Weights and biases initialization for calculating derivatives |
| innovationVector | Last hidden layer states' deltas update |
| layerEncoder | Layer encoder |
| loglik | Compute log-likelihood |
| meanA | Calculate mean of activated units |
| meanMz | Mean vector of units |
| meanVar | Mean, Jacobian and variance of activated units |
| meanVarDev | Mean and variance of activated units for derivatives |
| MedicalCost | Medical Cost of 1,338 insureds. |
| network | One iteration of the Tractable Approximate Gaussian Inference... |
| normalize | Normalize data |
| parameterBackwardPass | Backpropagation (parameters' deltas) |
| parameters | Indices for biases and weights |
| regression | Regression problem |
| runBatchDerivative | Result of the TAGI with derivative calculations |
| split | Split data |
| ToyExample.x_obs | Inputs used in training part for 1D toy problem |
| ToyExample.x_val | Inputs used in validation part for 1D toy problem |
| ToyExample.y_obs | Responses used in training part for 1D toy problem |
| ToyExample.y_val | Responses used in validation part for 1D toy problem |
