mbi.estimation.dual_averaging_step

mbi.estimation.dual_averaging_step(state: DualAveragingState, loss_fn: MarginalLossFn, marginal_oracle: MarginalOracle, total: Array | float, lipschitz: float, gamma: float, t: int) → DualAveragingState

Performs a single dual averaging step.

Parameters:

• state – Current algorithm state.

• loss_fn – The marginal loss function.

• marginal_oracle – A marginal oracle with signature (potentials, total) -> marginals.

• total – The known or estimated total number of records.

• lipschitz – Lipschitz constant of the gradient, divided by total.

• gamma – Variance-related parameter (typically 0 for deterministic).

• t – Current iteration number (1-indexed).

Returns:

Updated DualAveragingState.