ML (Training)

naive_bayes_training()¶

Demonstrates the naive_bayes_training function. This is the core FHE circuit logic used by FHENaiveBayesTrainer to calculate feature counts and class priors directly on encrypted data.

In [ ]:

from concrete import fhe
from concrete_fhe_toolkit.ml.training import naive_bayes_training

compiler = fhe.Compiler(naive_bayes_training, {'X_train': 'encrypted', 'y_train_one_hot': 'encrypted'})
inputset = [([[1, 0]], [[1, 0]])]
circuit = compiler.compile(inputset)

expected = naive_bayes_training(*inputset[0])
result = circuit.encrypt_run_decrypt(*inputset[0])
import numpy as np
np.testing.assert_array_equal(result[0], expected[0])
np.testing.assert_array_equal(result[1], expected[1])
print('naive_bayes_training compiled and executed successfully!')

from concrete import fhe from concrete_fhe_toolkit.ml.training import naive_bayes_training compiler = fhe.Compiler(naive_bayes_training, {'X_train': 'encrypted', 'y_train_one_hot': 'encrypted'}) inputset = [([[1, 0]], [[1, 0]])] circuit = compiler.compile(inputset) expected = naive_bayes_training(*inputset[0]) result = circuit.encrypt_run_decrypt(*inputset[0]) import numpy as np np.testing.assert_array_equal(result[0], expected[0]) np.testing.assert_array_equal(result[1], expected[1]) print('naive_bayes_training compiled and executed successfully!')

make_raw_naive_bayes_training()¶

Tests the function.

This cell verifies the make_raw_naive_bayes_training function mathematically against its cleartext counterpart, using a dynamically generated input set that covers positive, negative, zero, and array edge cases while respecting the [-7, 7] FHE RAM constraints.

In [ ]:

from concrete import fhe
from concrete_fhe_toolkit.ml.training import make_raw_naive_bayes_training

fn = make_raw_naive_bayes_training(thresholds=[1, 1])

def test_make_raw_naive_bayes_training_enc(X_train_raw, y_train_one_hot):
    import numpy as np
    res = fn(X_train_raw, y_train_one_hot)
    return fhe.array(res) if isinstance(res, list) else res

compiler = fhe.Compiler(test_make_raw_naive_bayes_training_enc, {'X_train_raw': 'encrypted', 'y_train_one_hot': 'encrypted'})
inputset = [([[1, 0]], [[1, 0]])]
circuit = compiler.compile(inputset)

_successes = 0
for inp in inputset:
    try:
        expected = fn(*inp)
        import numpy as np
        if isinstance(expected, (list, tuple)) or type(expected).__name__ == 'ndarray':
            np.testing.assert_array_equal(circuit.encrypt_run_decrypt(*inp), expected)
        else:
            assert int(circuit.encrypt_run_decrypt(*inp)) == int(expected), f"Failed at {inp}"
    except AssertionError:
        raise
    except Exception as e:
        print(f"Skipping {inp} due to bounds or other error: {e}")
    else:
        _successes += 1
assert _successes > 0, "make_raw_naive_bayes_training: all inputs were skipped — test is broken"
print(f"make_raw_naive_bayes_training tests passed! ({_successes}/{len(inputset)})")

from concrete import fhe from concrete_fhe_toolkit.ml.training import make_raw_naive_bayes_training fn = make_raw_naive_bayes_training(thresholds=[1, 1]) def test_make_raw_naive_bayes_training_enc(X_train_raw, y_train_one_hot): import numpy as np res = fn(X_train_raw, y_train_one_hot) return fhe.array(res) if isinstance(res, list) else res compiler = fhe.Compiler(test_make_raw_naive_bayes_training_enc, {'X_train_raw': 'encrypted', 'y_train_one_hot': 'encrypted'}) inputset = [([[1, 0]], [[1, 0]])] circuit = compiler.compile(inputset) _successes = 0 for inp in inputset: try: expected = fn(*inp) import numpy as np if isinstance(expected, (list, tuple)) or type(expected).__name__ == 'ndarray': np.testing.assert_array_equal(circuit.encrypt_run_decrypt(*inp), expected) else: assert int(circuit.encrypt_run_decrypt(*inp)) == int(expected), f"Failed at {inp}" except AssertionError: raise except Exception as e: print(f"Skipping {inp} due to bounds or other error: {e}") else: _successes += 1 assert _successes > 0, "make_raw_naive_bayes_training: all inputs were skipped — test is broken" print(f"make_raw_naive_bayes_training tests passed! ({_successes}/{len(inputset)})")