Functions — github.com/chritoth/active-bayesian-causal-inference

hub / github.com/chritoth/active-bayesian-causal-inference / functions

Functions218 in github.com/chritoth/active-bayesian-causal-inference

⨍ Functions 218◇ Types & classes 28

↓ 20 callersMethodclear_posterior_mll_cache

(self, keys: List[str] = None)

src/models/gp_model.py:96

↓ 17 callersMethodlog_prob

Returns the log probability for a given graph. Parameters ---------- graph : nx.DiGraph The query graph.

src/models/graph_models.py:175

↓ 17 callersMethodmll

(self, inputs: torch.Tensor, targets: torch.Tensor, prior_mode=False, reduce=True)

src/models/mechanisms.py:370

↓ 17 callersMethodsample

Generates samples a given input tensor according to the implemented likelihood model. Must be implemented by all child classes.

src/models/mechanisms.py:62

↓ 15 callersFunctionget_graph_key

Generates a unique string representation of a directed graph. Can be used as a dictionary key. Parameters ---------- graph : nx.DiGraph

src/models/graph_models.py:58

↓ 13 callersFunctionget_parents

Returns a list of parents for a given node in a given graph. Parameters ---------- node : str The child node. graph : nx.DiGr

src/models/graph_models.py:107

↓ 11 callersMethodsample

(self, interventions: dict, batch_size: int, num_batches: int = 1)

src/environments/environment.py:128

↓ 8 callersMethod__init__

(self, num_nodes: int, mechanism_model='gp-model', frac_non_intervenable_nod

src/environments/generic_environments.py:149

↓ 8 callersMethodcompute_importance_weights

(self, mc_graphs, use_cache=False, beta=50., log_weights: bool = False, dib

src/abci_dibs_gp.py:447

↓ 8 callersMethodsample_mc_graphs

(self, alpha: float = 1., set_data=False, num_graphs: int = None, only_dags=False)

src/abci_dibs_gp.py:339

↓ 7 callersMethod_check_args

Checks the generic argument shapes (inputs and targets) and their compatibility. Parameters ---------- inputs : torc

src/models/mechanisms.py:29

↓ 6 callersFunctiongather_data

(experiments: List[Experiment], node: str, graph: nx.DiGraph = None, parents: List[str] = None,

src/environments/experiment.py:19

↓ 6 callersMethodgraph_posterior_expectation

(self, func: Callable[[nx.DiGraph], torch.Tensor], logspace=False)

src/abci_categorical_gp.py:241

↓ 6 callersMethodgraph_posterior_expectation

(self, func: Callable[[nx.DiGraph], torch.Tensor], use_cache=True, logspace=False)

src/abci_dibs_gp.py:380

↓ 6 callersMethodselect_hyperparameters

(self, prior_mode: bool = False)

src/models/mechanisms.py:243

↓ 6 callersMethodset_data

(self, inputs: torch.Tensor, targets: torch.Tensor)

src/models/mechanisms.py:345

↓ 5 callersMethodentropy

Returns the entropy of the categorical distribution over graphs. Returns ---------- torch.Tensor The ent

src/models/graph_models.py:234

↓ 5 callersFunctionget_mechanism_key

(node, parents: List)

src/models/gp_model.py:13

↓ 5 callersMethodutility

(interventions: dict)

src/experimental_design/exp_designer_abci_dibs_gp.py:24

↓ 4 callersMethod__init__

Parameters ---------- in_size : int Number of mechanism inputs.

src/models/mechanisms.py:19

↓ 4 callersMethodedge_logits

(self, alpha: float = 1.)

src/models/graph_models.py:392

↓ 4 callersMethodedge_probs

(self, alpha: float = 1.)

src/models/graph_models.py:395

↓ 4 callersMethodeval

(self, keys: List[str] = None)

src/models/gp_model.py:78

↓ 4 callersMethodget_mc_graphs

Returns a set of graphs and corresponding log-weights for Monte Carlo estimation. Parameters ---------- mode: str

src/models/graph_models.py:300

↓ 4 callersMethodget_mechanism

(self, node, graph: nx.DiGraph = None, parents: List[str] = None)

src/models/gp_model.py:104

↓ 4 callersMethodget_random_intervention

(self, fixed_value: float = None)

src/abci_base.py:49

↓ 4 callersMethodget_target

(self, worker_id: int)

src/abci_base.py:67

↓ 4 callersMethodinit_design_process

(self, args: dict)

src/experimental_design/exp_designer_base.py:24

↓ 4 callersMethodlog_likelihood

(self, experiments: List[Experiment])

src/environments/environment.py:151

↓ 4 callersMethodmll

(self, experiments: List[Experiment], graph: nx.DiGraph, prior_mode=False, use_cache=False, mode='

src/models/gp_model.py:147

↓ 4 callersMethodquery_log_probs

(self, queries: List[InterventionalDistributionsQuery], graph: nx.DiGraph, num_imll_mc

src/models/gp_model.py:302

↓ 3 callersMethodclear_prior_mll_cache

(self, keys: List[str] = None)

src/models/gp_model.py:88

↓ 3 callersMethodcompute_graph_mlls

(self, graphs: List[List[nx.DiGraph]], experiments: List[Experiment] = None, prior_mode=True,

src/abci_dibs_gp.py:370

↓ 3 callersMethodcompute_graph_posterior

(self, experiments: List[Experiment], use_cache: bool = False)

src/abci_categorical_gp.py:230

↓ 3 callersMethodcompute_graph_posterior_mlls

(self, experiments: List[Experiment], graphs: List[List[nx.DiGraph]], mod

src/experimental_design/exp_designer_abci_dibs_gp.py:66

↓ 3 callersMethodcompute_graph_posterior_mlls

(self, experiments: List[Experiment], graphs: List[nx.DiGraph], mode='ind

src/experimental_design/exp_designer_abci_categorical_gp.py:49

↓ 3 callersMethodcreate_mechanism

(self, num_parents: int)

src/models/gp_model.py:217

↓ 3 callersMethoddesign_experiment

(self, target_node: str)

src/experimental_design/exp_designer_base.py:38

↓ 3 callersMethoddiscard_mechanisms

(self, current_time: int, max_age: int)

src/models/gp_model.py:64

↓ 3 callersFunctiongraph_to_adj_mat

Returns the adjecency matrix of the given graph as tensor. Parameters ---------- graph : nx.DiGraph The graph. node_labels :

src/models/graph_models.py:125

↓ 3 callersMethodload_param_dict

(self, param_dict)

src/environments/environment.py:221

↓ 3 callersMethodunnormalized_log_prior

(self, alpha: float = 1., beta: float = 1.)

src/models/graph_models.py:428

↓ 2 callersMethodadj_mat_to_graph

(self, adj_mat: torch.Tensor)

src/models/graph_models.py:523

↓ 2 callersFunctionauprc

(posterior_edge_probs: torch.Tensor, true_adj_mat: torch.Tensor)

src/utils/metrics.py:16

↓ 2 callersFunctionauroc

(posterior_edge_probs: torch.Tensor, true_adj_mat: torch.Tensor)

src/utils/metrics.py:7

↓ 2 callersMethodcompute_posterior_params

(self, targets: torch.Tensor, prior_mode=False)

src/models/mechanisms.py:94

↓ 2 callersMethodcreate_mechanism

(self, num_parents: int)

src/environments/environment.py:122

↓ 2 callersMethoddesign_experiment_distributed

(self, args)

src/abci_base.py:73

↓ 2 callersMethodexpected_noise_entropy

(self, prior_mode: bool = False)

src/models/mechanisms.py:389

↓ 2 callersMethodget_best_experiment

(self, target_nodes: Set[str])

src/experimental_design/exp_designer_base.py:62

↓ 2 callersMethodgraph_to_adj_mat

(self, graph: nx.DiGraph)

src/models/graph_models.py:529

↓ 2 callersMethodinit_graph_mechanisms

(self, graphs: List[List[nx.DiGraph]], set_data=False)

src/abci_dibs_gp.py:351

↓ 2 callersMethodinit_mechanisms

(self, graph: nx.DiGraph, init_time: int = 0)

src/models/gp_model.py:45

↓ 2 callersMethodload

(cls, path)

src/environments/environment.py:250

↓ 2 callersMethodlog_generative_prob

(self, adj_mats: torch.Tensor, alpha: float = 1., batch_mode=True)

src/models/graph_models.py:407

↓ 2 callersMethodprob

Returns the probability for a given graph. Parameters ---------- graph : nx.DiGraph The query graph.

src/models/graph_models.py:192

↓ 2 callersMethodreport_design

(self, worker_id: int, design_key: str, design: Design)

src/abci_base.py:60

↓ 2 callersMethodreport_status

(self, worker_id: int, message: str)

src/abci_base.py:64

↓ 2 callersFunctionresolve_mechanism_key

(key: str)

src/models/gp_model.py:19

↓ 2 callersMethodrun

(self, num_experiments=10, batch_size=1, update_interval=5, log_interval=5, num_initial_obs_samples=1)

src/abci_base.py:46

↓ 2 callersMethodsample

(self, interventions: dict, batch_size: int, num_batches: int, graph: nx.DiGraph)

src/models/gp_model.py:237

↓ 2 callersMethodsample

Samples a graph. Parameters ---------- num_graphs: int Number of graphs to sample. Returns

src/models/graph_models.py:246

↓ 2 callersMethodsample_initial_particles

(self, num_particles)

src/models/graph_models.py:384

↓ 2 callersMethodsample_intervention

(self)

src/environments/environment.py:19

↓ 2 callersMethodsample_queries

(self, queries: List[InterventionalDistributionsQuery], num_mc_queries: int, num_batche

src/models/gp_model.py:287

↓ 2 callersMethodsave

(self, path)

src/environments/environment.py:246

↓ 2 callersMethodset_data

(self, experiments: List[Experiment], keys: List[str] = None)

src/models/gp_model.py:220

↓ 2 callersMethodset_sample_queries

(self, sample_queries: List[Experiment])

src/environments/environment.py:22

↓ 2 callersFunctionshd

(target_graph: nx.DiGraph, predicted_graph: nx.DiGraph)

src/utils/metrics.py:25

↓ 2 callersFunctionspawn_abci_model

(abci_model, env, policy, num_workers)

src/scripts/run_single_env.py:18

↓ 2 callersMethodsubmodel

(self, graphs)

src/models/gp_model.py:384

↓ 2 callersMethodtrain

(self, keys: List[str] = None)

src/models/gp_model.py:83

↓ 2 callersMethodupdate_gp_hyperparameters

(self, update_time: int, experiments: List[Experiment], set_data=False, mech

src/models/gp_model.py:319

↓ 1 callersMethod_check_particle_shape

(self, z: torch.Tensor)

src/models/graph_models.py:381

↓ 1 callersMethodbackward

(ctx, grad_output)

src/models/graph_models.py:580

↓ 1 callersMethodclone

(self)

src/environments/environment.py:25

↓ 1 callersMethodcompute_graph_log_posterior

(self, graph: nx.DiGraph, alpha: float = 1.)

src/abci_dibs_gp.py:426

↓ 1 callersMethodcompute_posterior_edge_probs

(self, use_cache=True)

src/abci_dibs_gp.py:407

↓ 1 callersMethodconstruct_graph

(self, num_nodes: int)

src/environments/environment.py:197

↓ 1 callersMethoddagify_graphs

Uses a simple heuristic to 'dagify' cyclic graphs in-place. Note: this can be handy for testing and debugging during developement, but should

src/models/graph_models.py:470

↓ 1 callersMethoddraw_prior_hyperparams

(self)

src/models/mechanisms.py:240

↓ 1 callersMethoddraw_prior_hyperparams

(self)

src/models/mechanisms.py:289

↓ 1 callersMethodestimate_score_function

(self, alpha: float = 1., beta: float = 1.)

src/abci_dibs_gp.py:316

↓ 1 callersMethodexpected_cyclicity

(self, alpha: float = 1., num_samples: int = 100)

src/models/graph_models.py:435

↓ 1 callersMethodexperiment_designer_factory

(self)

src/abci_categorical_gp.py:33

↓ 1 callersMethodexperiment_designer_factory

(self)

src/abci_dibs_gp.py:39

↓ 1 callersMethodexperiment_designer_factory

(self)

src/abci_base.py:43

↓ 1 callersFunctiongenerate_all_dags

Generates all directed acyclic graphs with a given number of nodes or node labels. Parameters ---------- num_nodes : int The numb

src/models/graph_models.py:12

↓ 1 callersFunctiongenerate_job_id

()

src/scripts/run_single_env.py:40

↓ 1 callersFunctionget_free_port

()

src/scripts/run_single_env.py:34

↓ 1 callersMethodget_num_mechanisms

(self)

src/models/gp_model.py:174

↓ 1 callersMethodget_oracle_intervention

(self, num_samples: int, num_candidates_per_node: int = 10)

src/abci_categorical_gp.py:206

↓ 1 callersMethodget_parameters

(self, keys: List[str] = None)

src/models/gp_model.py:40

↓ 1 callersMethodget_result_files

(self, base_dir: str = '../results/')

src/utils/plotting.py:66

↓ 1 callersFunctiongp_ucb

(utility: callable, bounds: torch.Tensor, num_total_candidates=8, num_initial_candidates=1)

src/experimental_design/optimization.py:8

↓ 1 callersMethodgraph_info_gain

(self, interventions: dict, inner_mc_graphs: List[List[nx.DiGraph]],

src/experimental_design/exp_designer_abci_dibs_gp.py:75

↓ 1 callersMethodgraph_info_gain

(self, interventions: dict, batch_size: int = 1, num_exp_per_graph: int = 1, mode='n-best',

src/experimental_design/exp_designer_abci_categorical_gp.py:56

↓ 1 callersFunctiongrid_search

(utility: callable, bounds: torch.Tensor, num_candidates=10)

src/experimental_design/optimization.py:62

↓ 1 callersMethodhyperparam_log_prior

(self, prior_mode: bool = False)

src/models/mechanisms.py:234

↓ 1 callersMethodinit_as_static

(self)

src/models/mechanisms.py:112

next →1–100 of 218, ranked by callers