Homoscedastic Uncertainty

class softsensor.homoscedastic_model.HomoscedasticModel(model, homoscedastic_var)

Wrapper class for point prediction models that uses a constant variance for all inputs

Parameters:

• model (uncertainty model) – model that is used for prediction

• homoscedastic_var (tensor[float], shape=[model.pred_size]) – Homoscedastic variance

Return type:

None.

Examples

Define Model

>>> import softsensor.autoreg_models
>>> import softsensor.homoscedastic_model as hm
>>> import torch
>>> m = softsensor.autoreg_models.ARNN(2, 1, 10, 10, [16, 8])
>>> vars = torch.tensor([1])
>>> homosc_m = hm.HomoscedasticModel(m, vars)
>>> input = torch.randn(32, 2, 10)
>>> rec_input = torch.randn(32, 1, 10)
>>> output = homosc_m(input, rec_input)
>>> print(output[0].shape)
torch.Size([32, 1, 1])
>>> print(output[1].shape)
torch.Size([32, 1, 1])

Define Data

>>> import softsensor.meas_handling as ms
>>> import numpy as np
>>> import pandas as pd
>>> t = np.linspace(0, 1.0, 101)
>>> d = {'inp1': np.random.randn(101),
         'inp2': np.random.randn(101),
         'out': np.random.randn(101)}
>>> handler = ms.Meas_handling([pd.DataFrame(d, index=t)], ['train'],
                               ['inp1', 'inp2'], ['out'], fs=100)

Define Model with Uncertainty

>>> loader = handler.give_list(window_size=10, keyword='training',
                               rnn_window=10, batch_size=1)
>>> vars = torch.tensor([1])
>>> homosc_m = hm.HomoscedasticModel(m, vars)
>>> mean, var = homosc_m.prediction(loader[0])
>>> print(mean.shape)
torch.Size([1, 101])
>>> print(var.shape)
torch.Size([1, 101])

estimate_uncertainty_mean_std(*args)

forward(*args)

Forward function to propagate through the network

Parameters:

*args (input arguments) – (need to be the same as the Point model in __init__() needs)

Returns:

• mean (torch.tensor dtype=torch.float()) – shape=[batch size, pred_size]

• std (torch.tensor dtype=torch.float() in [0,1]) – shape=[batch size, pred_size]

prediction(dataloader, device='cpu')

Prediction of a whole Time Series

Parameters:

• dataloader (Dataloader) – Dataloader to predict output

• device (str, optional) – device to compute on. The default is ‘cpu’.

Returns:

• if loss_ft=None –

  (torch.Tensor, list[torch.Tensor])

  tuple of Torch Tensor of same length as input and var

• if loss_ft=torch loss funciton –

  (torch.Tensor, list[torch.Tensor], loss)

  tuple of Torch Tensor of same length as input, var and loss

softsensor.homoscedastic_model.fit_homoscedastic_var(dataframes, model_out, target_out)

Determines a homoscedastic uncertainty as the average sample variance on the training set where each point prediction is regarded as a single sample from a distribution with the ground truth as mean.

Parameters:

• model (torch.Module) – Point prediction model

• train_list (list[Dataloader]) – dataset to fit the homoscedastic variance

Returns:

homoscedastic_var – Homoscedastic variance

Return type:

tensor[float], shape=[model.pred_size]