Skip to content

Cumulitive update #49

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Maslyaev merged 7 commits into from
Oct 13, 2025
Merged

Cumulitive update #49

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
66cfd48
removed inner weights adjustment
Gromwud Oct 13, 2025
f1eaaca
l2 normalization of objective functions and windows step bugfix
Gromwud Oct 13, 2025
5d7c4ec
l2 normalization of features for the LASSO and right part selection f…
Gromwud Oct 13, 2025
60a5fa8
PopulationUpdater bugfix and OffspringUpdater update
Gromwud Oct 13, 2025
b83fb0c
right part selection logic updated
Gromwud Oct 13, 2025
4f4631f
mutation update
Gromwud Oct 13, 2025
5a54e6e
Merge remote-tracking branch 'origin/main'
Gromwud Oct 13, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 3 additions & 3 deletions epde/operators/common/coeff_calculation.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -82,9 +82,9 @@ def apply(self, objective : Equation, arguments : dict = None):
if weight_idx in nonzero_features_indexes:
weights[weight_idx] = valueable_weights[nonzero_features_indexes.index(weight_idx)]
weights[-1] = valueable_weights[-1]
nonzero_terms_mask = np.array([False if np.isclose(weight, 0) else True for weight in weights])
weights = np.array([item if keep else 0 for item, keep in zip(weights, nonzero_terms_mask)])
objective.weights_internal = np.array([item if keep else 0 for item, keep in zip(objective.weights_internal, nonzero_terms_mask[:-1])])
# nonzero_terms_mask = np.array([False if np.isclose(weight, 0) else True for weight in weights])
# weights = np.array([item if keep else 0 for item, keep in zip(weights, nonzero_terms_mask)])
# objective.weights_internal = np.array([item if keep else 0 for item, keep in zip(objective.weights_internal, nonzero_terms_mask[:-1])])
objective.weights_final_evald = True
objective.weights_final = weights

Expand Down
64 changes: 18 additions & 46 deletions epde/operators/common/fitness.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -127,15 +127,11 @@ def apply(self, objective: Equation, arguments: dict, force_out_of_place: bool =

try:
if features is None:
maximum = target.max(axis=0)
minimum = target.min(axis=0)
discr = (target - target.mean(axis=0) - minimum) / (maximum - minimum)
discr = (target - target.mean(axis=0)) / np.linalg.norm(target, 2)
else:
discr_feats = np.dot(features, objective.weights_final[:-1][objective.weights_internal != 0])
discr_feats = discr_feats + np.full(target.shape, objective.weights_final[-1])
maximum = np.max([discr_feats.max(axis=0), target.max(axis=0)])
minimum = np.min([discr_feats.min(axis=0), target.min(axis=0)])
discr = ((discr_feats - minimum) - (target - minimum)) / (maximum - minimum)
discr = (discr_feats - target) / np.linalg.norm(target, 2)
discr = np.multiply(discr, self.g_fun_vals)
rl_error = np.linalg.norm(discr, ord=2)
except ValueError:
Expand All @@ -145,32 +141,14 @@ def apply(self, objective: Equation, arguments: dict, force_out_of_place: bool =
raise ValueError('Incorrect penalty coefficient set, value shall be in (0, 1).')

fitness_value = rl_error
# if np.sum(objective.weights_final) == 0:
# fitness_value /= self.params['penalty_coeff']

if force_out_of_place:
return fitness_value

# discr = np.mean(discr ** 2)
# ll = np.log(discr)
# aic = 2 * len(objective.weights_final) - 2 * ll
# ssr = np.sum(discr ** 2)
# n = len(target)
# llf = - n / 2 * np.log(2 * np.pi) - n / 2 * np.log(ssr / n) - n / 2
# aic = 2 * len([_ for _ in objective.weights_final if _ != 0]) - 2 * llf
# aic = np.log(n) * len([_ for _ in objective.weights_final if _ != 0]) - 2 * llf
# objective.aic = 1/(1 + np.exp(- 1e-4 * ll))

# if force_out_of_place:
# return 1 / (np.exp(-aic / 3e5))

# objective.aic = 1 / (np.exp(-aic / 3e5))
objective.aic = None
objective.aic_calculated = True
# print(aic)
# print(len([_ for _ in objective.weights_final if _ !=0]))
# print(objective.aic)
assert objective.simplified, 'Trying to evaluate not simplified equation.'

# assert objective.simplified, 'Trying to evaluate not simplified equation.'

# Calculate r-loss
data_shape = global_var.grid_cache.g_func.shape
Expand All @@ -190,7 +168,7 @@ def apply(self, objective: Equation, arguments: dict, force_out_of_place: bool =
if target_vals.ndim == 1:
window_size = len(target_vals) // 2
num_horizons = len(target_vals) - window_size + 1
if window_size < 15:
if num_horizons < 30:
step_size = 1
else:
step_size = num_horizons // 30
Expand All @@ -200,10 +178,10 @@ def apply(self, objective: Equation, arguments: dict, force_out_of_place: bool =
for start_idx in range(0, num_horizons, step_size):
end_idx = start_idx + window_size
target_window = target_vals[start_idx:end_idx]
if np.isclose(np.sqrt(np.mean(np.power(target_window, 2))), 0, atol=1e-10):
if np.isclose(np.linalg.norm(target_window, 2), 0):
window_stability = np.abs(np.std(target_window))
else:
window_stability = np.abs(np.std(target_window) / np.sqrt(np.mean(np.power(target_window, 2))))
window_stability = np.abs(np.std(target_window) / np.linalg.norm(target_window, 2))
eq_window_weights.append(window_stability)
lr = np.mean(eq_window_weights)
else:
Expand All @@ -217,8 +195,8 @@ def apply(self, objective: Equation, arguments: dict, force_out_of_place: bool =
valuable_weights = estimator.coef_[:-1]
eq_window_weights.append(valuable_weights)
eq_cv = np.array([
np.abs(np.std(_)) if np.isclose(np.sqrt(np.mean(np.power(_, 2))), 0, atol=1e-10)
else np.abs(np.std(_) / np.sqrt(np.mean(np.power(_, 2))))
np.abs(np.std(_)) if np.isclose(np.linalg.norm(_, 2), 0)
else np.abs(np.std(_) / np.linalg.norm(_, 2))
for _ in zip(*eq_window_weights)
])
lr = eq_cv.mean()
Expand All @@ -229,7 +207,7 @@ def apply(self, objective: Equation, arguments: dict, force_out_of_place: bool =
eq_window_weights = []
window_size = target_vals.shape[dim] // 2
num_horizons = target_vals.shape[dim] - window_size + 1
if window_size < 15:
if num_horizons < 30:
step_size = 1
else:
step_size = num_horizons // 30
Expand All @@ -241,10 +219,10 @@ def apply(self, objective: Equation, arguments: dict, force_out_of_place: bool =
target_window = target_vals[start_idx:end_idx, :].reshape(-1)
else:
target_window = target_vals[:, start_idx:end_idx].reshape(-1)
if np.isclose(np.sqrt(np.mean(np.power(target_window, 2))), 0, atol=1e-10):
if np.isclose(np.linalg.norm(target_window, 2), 0):
window_stability = np.abs(np.std(target_window))
else:
window_stability = np.abs(np.std(target_window) / np.sqrt(np.mean(np.power(target_window, 2))))
window_stability = np.abs(np.std(target_window) / np.linalg.norm(target_window, 2))
eq_window_weights.append(window_stability)
lr += np.mean(eq_window_weights)
else:
Expand All @@ -263,8 +241,8 @@ def apply(self, objective: Equation, arguments: dict, force_out_of_place: bool =
valuable_weights = estimator.coef_[:-1]
eq_window_weights.append(valuable_weights)
eq_cv = np.array([
np.abs(np.std(_)) if np.isclose(np.sqrt(np.mean(np.power(_, 2))), 0, atol=1e-10)
else np.abs(np.std(_) / np.sqrt(np.mean(np.power(_, 2))))
np.abs(np.std(_)) if np.isclose(np.linalg.norm(_, 2), 0)
else np.abs(np.std(_) / np.linalg.norm(_, 2))
for _ in zip(*eq_window_weights)
])
lr += eq_cv.mean()
Expand All @@ -285,10 +263,10 @@ def apply(self, objective: Equation, arguments: dict, force_out_of_place: bool =
target_window = target_vals[:, start_idx:end_idx, :].reshape(-1)
else:
target_window = target_vals[:, :, start_idx:end_idx].reshape(-1)
if np.isclose(np.sqrt(np.mean(np.power(target_window, 2))), 0, atol=1e-10):
if np.isclose(np.linalg.norm(target_window, 2), 0):
window_stability = np.abs(np.std(target_window))
else:
window_stability = np.abs(np.std(target_window) / np.sqrt(np.mean(np.power(target_window, 2))))
window_stability = np.abs(np.std(target_window) / np.linalg.norm(target_window, 2))
eq_window_weights.append(window_stability)
lr += np.mean(eq_window_weights)
else:
Expand All @@ -311,21 +289,15 @@ def apply(self, objective: Equation, arguments: dict, force_out_of_place: bool =
valuable_weights = estimator.coef_[:-1]
eq_window_weights.append(valuable_weights)
eq_cv = np.array([
np.abs(np.std(_)) if np.isclose(np.sqrt(np.mean(np.power(_, 2))), 0, atol=1e-10)
else np.abs(np.std(_) / np.sqrt(np.mean(np.power(_, 2))))
np.abs(np.std(_)) if np.isclose(np.linalg.norm(_, 2), 0)
else np.abs(np.std(_) / np.linalg.norm(_, 2))
for _ in zip(*eq_window_weights)
])

lr += eq_cv.mean()

fitness_value = round(fitness_value, 8)
lr = round(lr / target_vals.ndim, 8)

# if lr > self.params['max_lr']:
# self.params['max_lr'] = lr
# if fitness_value > self.params['max_lp']:
# self.params['max_lp'] = fitness_value

objective.fitness_calculated = True
objective.fitness_value = fitness_value
objective.stability_calculated = True
Expand Down
11 changes: 6 additions & 5 deletions epde/operators/common/right_part_selection.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -48,7 +48,8 @@ def apply(self, objective : Equation, arguments : dict):

objective.reset_state(True)

while not (objective.right_part_selected and objective.simplified):
while not (objective.simplified and objective.is_correct_right_part):
objective.is_correct_right_part = False
min_fitness = np.inf
weights_internal = np.zeros_like(objective.structure)
min_idx = 0
Expand All @@ -75,8 +76,11 @@ def apply(self, objective : Equation, arguments : dict):
# if not np.isclose(objective.fitness_value, max_fitness) and global_var.verbose.show_warnings:
# warnings.warn('Reevaluation of fitness function for equation has obtained different result. Not an error, if ANN DE solver is used.')
self.simplify_equation(objective)
if objective.structure[objective.target_idx].contains_variable(objective.main_var_to_explain) and objective.structure[objective.target_idx].contains_deriv(objective.main_var_to_explain):
objective.is_correct_right_part = True
else:
objective.reset_explaining_term(objective.target_idx)
objective.right_part_selected = True

def simplify_equation(self, objective: Equation):
# Get nonzero terms
Expand Down Expand Up @@ -121,11 +125,8 @@ def simplify_equation(self, objective: Equation):
while objective.structure.count(term) > 1 or term == temp:
term.randomize()
term.reset_saved_state()
objective.simplified = False
objective.right_part_selected = False
return
return
objective.simplified = True
objective.right_part_selected = True

def use_default_tags(self):
self._tags = {'equation right part selection', 'gene level', 'contains suboperators', 'inplace'}
Expand Down
Loading