gw.transforms.gnpe_transforms.GNPEBase
gw.transforms.gnpe_transforms.GNPEBase(kernel_dict, operators)A base class for Group Equivariant Neural Posterior Estimation 1.
This implements GNPE for approximate equivariances. For exact equivariances, additional processing should be implemented within a subclass.
Attributes
| Name | Description |
|---|---|
| context_parameters | |
| input_parameter_names | |
| kernel | |
| operators | |
| proxy_list |
Methods
| Name | Description |
|---|---|
| inverse | |
| multiply | |
| perturb | Generate proxy variables based on initial parameter values. |
| sample_proxies | Given input parameters, perturbs based on the |
inverse
gw.transforms.gnpe_transforms.GNPEBase.inverse(a, k)multiply
gw.transforms.gnpe_transforms.GNPEBase.multiply(a, b, k)perturb
gw.transforms.gnpe_transforms.GNPEBase.perturb(g, k)Generate proxy variables based on initial parameter values.
Parameters
| Name | Type | Description | Default |
|---|---|---|---|
| g | Union[np.float64, float, torch.Tensor] | Initial parameter values | required |
| k | str | Parameter name. This is used to identify the group binary operator. | required |
Returns
| Name | Type | Description |
|---|---|---|
| Proxy variables in the same format as g. |
sample_proxies
gw.transforms.gnpe_transforms.GNPEBase.sample_proxies(input_parameters)Given input parameters, perturbs based on the kernel to produce “proxy” (“hatted”) parameters, i.e., samples
\hat g ~ p(\hat g | g).Typically the GNPE NDE will be conditioned on g. Furthermore, these proxy parameters will be used to transform the data to simplify it.
Parameters:
input_parameters : dict Initial parameter values to be perturbed. dict values can be either floats (for training) or torch Tensors (for inference).
Returns
| Name | Type | Description |
|---|---|---|
| A dict of proxy parameters. |