uqregressors.bayesian.deep_ens
Deep Ensembles are implemented as in Lakshimnarayanan et al. 2017.
Deep Ensembles
This module implements Deep Ensemble Regressors for regression of a one dimensional output.
Key features are
- Customizable neural network architecture
- Prediction Intervals based on Gaussian assumption
- Parallel training of ensemble members with Joblib
- Customizable optimizer and loss function
- Optional Input/Output Normalization
DeepEnsembleRegressor
Bases: BaseEstimator
, RegressorMixin
Deep Ensemble Regressor with uncertainty estimation using neural networks.
Trains an ensemble of MLP models to predict both mean and variance for regression tasks, and provides predictive uncertainty intervals.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
name
|
str
|
Name of the regressor for config files. |
'Deep_Ensemble_Regressor'
|
n_estimators
|
int
|
Number of ensemble members. |
5
|
hidden_sizes
|
list of int
|
List of hidden layer sizes for each MLP. |
[64, 64]
|
alpha
|
float
|
Significance level for prediction intervals (e.g., 0.1 for 90% interval). |
0.1
|
requires_grad
|
bool
|
If True, returned predictions require gradients. |
False
|
activation_str
|
str
|
Name of activation function to use (e.g., 'ReLU'). |
'ReLU'
|
learning_rate
|
float
|
Learning rate for optimizer. |
0.001
|
epochs
|
int
|
Number of training epochs. |
200
|
batch_size
|
int
|
Batch size for training. |
32
|
optimizer_cls
|
Optimizer
|
Optimizer class. |
Adam
|
optimizer_kwargs
|
dict
|
Additional kwargs for optimizer. |
None
|
scheduler_cls
|
_LRScheduler or None
|
Learning rate scheduler class. |
None
|
scheduler_kwargs
|
dict
|
Additional kwargs for scheduler. |
None
|
loss_fn
|
callable
|
Loss function accepting (preds, targets). |
None
|
device
|
str or device
|
Device to run training on ('cpu' or 'cuda'). |
'cpu'
|
use_wandb
|
bool
|
Whether to use Weights & Biases logging. |
False
|
wandb_project
|
str or None
|
WandB project name. |
None
|
wandb_run_name
|
str or None
|
WandB run name prefix. |
None
|
n_jobs
|
int
|
Number of parallel jobs to train ensemble members. |
1
|
random_seed
|
int or None
|
Seed for reproducibility. |
None
|
scale_data
|
bool
|
Whether to scale input and output data. |
True
|
input_scaler
|
object or None
|
Scaler for input features. |
None
|
output_scaler
|
object or None
|
Scaler for target values. |
None
|
tuning_loggers
|
list
|
List of tuning loggers. |
[]
|
Attributes:
Name | Type | Description |
---|---|---|
models |
list
|
List of trained PyTorch MLP models. |
input_dim |
int
|
Dimensionality of input features. |
_loggers |
list
|
Training loggers for each model. |
training_logs |
Logs from training. |
|
tuning_logs |
Logs from hyperparameter tuning. |
Source code in uqregressors\bayesian\deep_ens.py
63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 |
|
fit(X, y)
Fit the ensemble on training data.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
X
|
array - like or Tensor
|
Training inputs. |
required |
y
|
array - like or Tensor
|
Training targets. |
required |
Returns:
Type | Description |
---|---|
DeepEnsembleRegressor
|
Fitted estimator. |
Source code in uqregressors\bayesian\deep_ens.py
245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 |
|
load(path, device='cpu', load_logs=False)
classmethod
Load a saved ensemble regressor from disk.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
path
|
str or Path
|
Directory path to load the model from. |
required |
device
|
str or device
|
Device to load the model onto. |
'cpu'
|
load_logs
|
bool
|
Whether to load training and tuning logs. |
False
|
Returns:
Type | Description |
---|---|
DeepEnsembleRegressor
|
Loaded model instance. |
Source code in uqregressors\bayesian\deep_ens.py
367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 |
|
nll_loss(preds, y)
Negative log-likelihood loss assuming Gaussian outputs.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
preds
|
Tensor
|
Predicted means and variances, shape (batch_size, 2). |
required |
y
|
Tensor
|
True target values, shape (batch_size,). |
required |
Returns:
Type | Description |
---|---|
Tensor
|
Scalar loss value. |
Source code in uqregressors\bayesian\deep_ens.py
166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 |
|
predict(X)
Predicts the target values with uncertainty estimates.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
X
|
ndarray
|
Feature matrix of shape (n_samples, n_features). |
required |
Returns:
Type | Description |
---|---|
Union[Tuple[ndarray, ndarray, ndarray], Tuple[Tensor, Tensor, Tensor]]
|
Tuple containing: mean predictions, lower bound of the prediction interval, upper bound of the prediction interval. |
Note
If requires_grad
is False, all returned arrays are NumPy arrays.
Otherwise, they are PyTorch tensors with gradients.
Source code in uqregressors\bayesian\deep_ens.py
273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 |
|
save(path)
Save the trained ensemble to disk.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
path
|
str or Path
|
Directory path to save the model and metadata. |
required |
Source code in uqregressors\bayesian\deep_ens.py
326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 |
|
MLP
Bases: Module
A simple multi-layer perceptron which outputs a mean and a positive variance per input sample.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
input_dim
|
int
|
Number of input features. |
required |
hidden_sizes
|
list of int
|
List of hidden layer sizes. |
required |
activation
|
Module
|
Activation function class (e.g., nn.ReLU). |
required |
Source code in uqregressors\bayesian\deep_ens.py
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 |
|