def test_line():
model = line_potts_model(gr_size=40, al_size=3, seed=0)
gt = model.infer(algorithm='path_dp')
result = model.infer(algorithm='mean_field')
assert_results_close(result, gt, log_pf_tol=3.0, mp_mse_tol=1e-2)
def test_grid():
model = grid_potts_model(4, 5, al_size=2, seed=0)
gt = model.infer(algorithm='path_dp')
result = model.infer(algorithm='mean_field', max_iter=1000)
assert_results_close(result, gt, log_pf_tol=5.0, mp_mse_tol=0.05)
def test_tree():
model = tree_potts_model(gr_size=50, al_size=3, seed=0)
gt = model.infer(algorithm='tree_dp')
result = model.infer(algorithm='mean_field')
assert_results_close(result, gt, log_pf_tol=10.0, mp_mse_tol=0.03)
def test_clique():
model = clique_potts_model(gr_size=5, al_size=3, seed=0)
gt = model.infer(algorithm='bruteforce')
result = model.infer(algorithm='mean_field')
assert_results_close(result, gt, log_pf_tol=1.0, mp_mse_tol=1e-4)
def test_small_forest():
edges = [[0, 1], [1, 2], [1, 3], [4, 5], [6, 7]]
field = np.random.random((8, 2))
inter = np.random.random((5, 2, 2))
model = PairWiseFiniteModel.create(field, edges, inter)
assert_results_close(model.infer(algorithm='message_passing'),
model.infer(algorithm='bruteforce'))
def test_clique():
model = clique_potts_model(gr_size=5, al_size=3, seed=0)
gt = infer_bruteforce(model)
result = model.infer(algorithm='path_dp')
assert_results_close(result, gt)
def test_marg_probs_potts_tree_1000x5():
libdai = LibDaiInterop()
if not libdai.is_libdai_ready():
return
model = tree_potts_model(1000, 5)
true_result = model.infer()
libdai_result = libdai.infer(model, "BP")
assert_results_close(true_result, libdai_result, log_pf_tol=1e-8)
def test_grid():
model = grid_potts_model(10, 5, al_size=2)
true_result = model.infer(algorithm='path_dp')
result_bp = BP.infer(model)
assert_results_close(true_result,
result_bp,
log_pf_tol=0.1,
mp_mse_tol=2e-3)
def test_long_line_compare_with_tree():
# Test to ensure Tree DP and Path DP are consistent.
gr_size = 1000
al_size = 3
model = line_potts_model(gr_size=gr_size, al_size=al_size, seed=111)
result1 = model.infer(algorithm='path_dp')
result2 = model.infer(algorithm='tree_dp')
assert_results_close(result1, result2)
def test_libdai_bp_regression():
model = random_generic_model(num_variables=20,
num_factors=20,
max_domain_size=4,
max_factor_size=3)
default_opts = {
'tol': 1e-9,
'logdomain': 0,
'updates': 'SEQFIX',
'verbose': 0
}
configs = [
{
**default_opts
},
{
**default_opts, 'logdomain': 1
},
{
**default_opts, 'updates': 'SEQMAX'
},
{
**default_opts, 'updates': 'PARALL'
},
{
**default_opts, 'updates': 'SEQRND'
},
{
**default_opts, 'maxiter': 0
},
{
**default_opts, 'maxiter': 1
},
{
**default_opts, 'maxiter': 5,
'logdomain': 1
},
{
**default_opts, 'maxiter': 50
},
{
**default_opts, 'damping': 0.1,
'logdomain': 0,
'maxiter': 5
},
{
**default_opts, 'damping': 0.2,
'logdomain': 1,
'maxiter': 10
},
]
for options in configs:
result_libdai = libdai.infer(model, algorithm='BP', options=options)
result_bp = BP.infer(model, options)
assert_results_close(result_libdai, result_bp)
def test_cycle3():
np.random.seed(0)
model = PairWiseFiniteModel(3, 3)
for i, j in [(0, 1), (1, 2), (0, 2)]:
model.add_interaction(i, j, np.random.random(size=(3, 3)))
gt = model.infer(algorithm='bruteforce')
result = model.infer(algorithm='mean_field')
assert_results_close(result, gt, log_pf_tol=0.1, mp_mse_tol=1e-4)
def test_isolated_exact():
np.random.seed(0)
gr_size = 1000
al_size = 5
model = PairWiseFiniteModel(gr_size, al_size)
model.set_field(np.random.random((gr_size, al_size)))
gt = model.infer(algorithm='bruteforce')
result = model.infer(algorithm='mean_field')
assert_results_close(result, gt)
def test_random_forest_compare_with_tree_dp():
# Generate tree on 50 edges, then leave only 40 random edges.
gr_size, al_size = 50, 5
edges = list(networkx.random_tree(gr_size).edges())
random.shuffle(edges)
edges = edges[0:40]
model = PairWiseFiniteModel(gr_size, 5)
model.set_field(np.random.random((gr_size, al_size)))
for v1, v2 in edges:
model.add_interaction(v1, v2, np.random.random((al_size, al_size)))
assert_results_close(model.infer(algorithm='message_passing'),
model.infer(algorithm='tree_dp'))
def test_inference_all_methods():
# Sanity check that all algorithms work on very simple model.
all_methods = [
'auto', 'bruteforce', 'mean_field', 'message_passing', 'tree_dp',
'path_dp', 'junction_tree'
]
model = PairWiseFiniteModel(2, 2)
model.add_interaction(0, 1, np.array([[0, 0], [0, 1]]))
m = np.array([2, 1 + np.exp(1)])
expected_result = InferenceResult(np.log(3 + np.exp(1)),
np.array([m, m]) / np.sum(m))
for method in all_methods:
result = model.infer(algorithm=method)
assert_results_close(result,
expected_result,
log_pf_tol=1.0,
mp_mse_tol=0.1)
def test_inference_tree_100x5():
model = tree_potts_model(gr_size=100, al_size=5, seed=0)
ground_truth = model.infer(algorithm='tree_dp')
result = infer_junction_tree(model)
assert_results_close(result, ground_truth)
def test_grid_3x3():
model = grid_potts_model(3, 3, al_size=2, seed=123)
gt = model.infer(algorithm='bruteforce')
result = model.infer(algorithm='path_dp')
assert_results_close(result, gt)
def test_tree_exact():
model = tree_potts_model(gr_size=50, al_size=3, seed=0)
assert_results_close(model.infer(algorithm='message_passing'),
model.infer(algorithm='tree_dp'))
def test_tree():
model = tree_potts_model(gr_size=100, al_size=3)
true_result = model.infer(algorithm='tree_dp')
result_bp = BP.infer(model)
assert_results_close(true_result, result_bp, log_pf_tol=2e-9)
def test_fully_isolated():
model = PairWiseFiniteModel(10, 2)
model.set_field(np.random.random(size=(10, 2)))
ground_truth = model.infer(algorithm='bruteforce')
result = model.infer(algorithm='tree_dp')
assert_results_close(result, ground_truth)
def test_vert15_alph2():
model = tree_potts_model(gr_size=10, al_size=2, seed=123)
ground_truth = model.infer(algorithm='bruteforce')
result = model.infer(algorithm='tree_dp')
assert_results_close(result, ground_truth)
def test_potts_grid():
model = grid_potts_model(6, 5, al_size=3)
true_result = model.infer(algo='parth_dp')
model = inferlo.GenericGraphModel.from_model(model)
bw_result = BackwardBucketElimination.infer(model)
assert_results_close(true_result, bw_result)
def test_inference_clique_10x2():
model = clique_potts_model(gr_size=10, al_size=2, seed=0)
ground_truth = model.infer(algorithm='bruteforce')
result = infer_junction_tree(model)
assert_results_close(result, ground_truth)
def test_inference_cross_50x2x2():
model = cross_potts_model(length=50, width=2, al_size=2)
ground_truth = model.infer(algorithm='path_dp')
result = infer_junction_tree(model)
assert_results_close(result, ground_truth)
def test_inference_grid_4x50x2():
model = grid_potts_model(4, 50, al_size=2, seed=0)
ground_truth = model.infer(algorithm='path_dp')
result = infer_junction_tree(model)
assert_results_close(result, ground_truth)
