omnibias.boolean.jax.ops.design
Boolean truth tables, multilinear extensions, algebraic spectra, and finite-temperature gate systems.
import omnibias.boolean.jax.ops.designModule contract & mathematical conventions
Differentiable Boolean-function design losses (jax).
Twin of omnibias.boolean.torch.ops.design: spectral objectives
(degree_penalty, influence_penalty, target_spectrum_loss)
built on the differentiable ~omnibias.boolean.jax.ops.spectrum engine.
Shared formulas do not imply universal floating-point identity. Verify dtype, shape, derivative-order support, and the stated numerical assumptions for the path you use.
degree_penalty
omnibias.boolean.jax.ops.degree_penaltySource ↗degree_penalty(values: Sequence[float] | Array) -> Arraysum_S |S| * hat f(S)^2 -- spectral energy weighted by monomial order.
Parameters
| Name | Type & default | Contract |
|---|---|---|
valuespositional-or-keyword | Sequence[float] | ArrayRequired | See the operation contract and implementation below. |
ArrayExecuted regression · design loss parity
Recorded coverage confirms this test executed the API body. This example belongs to the pinned upstream test suite. It may use fixtures and imports from its full file. Run it from that checkout with its test dependencies installed.
def test_design_loss_parity() -> None:
vals = [0.1, 0.7, 0.3, 0.9]
t = float(tb.degree_penalty(torch.tensor(vals, dtype=torch.float64)))
j = float(jb.degree_penalty(jnp.asarray(vals, dtype=jnp.float64)))
assert t == pytest.approx(j, rel=1e-9, abs=1e-11)python -m pytest "packages/omnibias-boolean/tests/test_cross_backend.py::test_design_loss_parity" -qRead the implementation
def degree_penalty(values: Sequence[float] | Array) -> Array:
"""``sum_S |S| * hat f(S)^2`` -- spectral energy weighted by monomial order."""
w = walsh_coeffs(values)
n = w.shape[0].bit_length() - 1
orders = jnp.asarray(
[bin(mask).count("1") for mask in range(1 << n)], dtype=w.dtype
)
return (orders * w**2).sum()Other public import paths
omnibias.boolean.jax.ops.degree_penaltyomnibias.boolean.jax.ops.design.degree_penalty
influence_penalty
omnibias.boolean.jax.ops.influence_penaltySource ↗influence_penalty(values: Sequence[float] | Array) -> ArrayTotal influence sum_i Inf_i (average sensitivity), differentiable.
Parameters
| Name | Type & default | Contract |
|---|---|---|
valuespositional-or-keyword | Sequence[float] | ArrayRequired | See the operation contract and implementation below. |
ArrayRead the implementation
def influence_penalty(values: Sequence[float] | Array) -> Array:
"""Total influence ``sum_i Inf_i`` (average sensitivity), differentiable."""
return influences_diff(values).sum()Other public import paths
omnibias.boolean.jax.ops.influence_penaltyomnibias.boolean.jax.ops.design.influence_penalty
target_spectrum_loss
omnibias.boolean.jax.ops.target_spectrum_lossSource ↗target_spectrum_loss(values: Sequence[float] | Array, target: Sequence[float] | Array, basis: str='walsh') -> ArrayMean-squared error between the function's spectrum and a target spectrum.
Parameters
| Name | Type & default | Contract |
|---|---|---|
valuespositional-or-keyword | Sequence[float] | ArrayRequired | See the operation contract and implementation below. |
targetpositional-or-keyword | Sequence[float] | ArrayRequired | See the operation contract and implementation below. |
basispositional-or-keyword | strDefault: 'walsh' | See the operation contract and implementation below. |
ArrayRead the implementation
def target_spectrum_loss(
values: Sequence[float] | Array,
target: Sequence[float] | Array,
basis: str = "walsh",
) -> Array:
"""Mean-squared error between the function's spectrum and a target spectrum."""
if basis == "walsh":
coeffs = walsh_coeffs(values)
elif basis == "mobius":
coeffs = mobius_coeffs(values)
else:
raise ValueError(f"basis must be 'walsh' or 'mobius', got {basis!r}")
target_a = jnp.asarray(target, dtype=coeffs.dtype)
return ((coeffs - target_a) ** 2).mean()Other public import paths
omnibias.boolean.jax.ops.target_spectrum_lossomnibias.boolean.jax.ops.design.target_spectrum_loss