Skip to main content

finite_difference_gradient

tensorlogic_sklears_kernels::deep_kernel::gradient

Function finite_difference_gradient 

Source 
pub fn finite_difference_gradient<K: Kernel>(
    kernel: &mut DeepKernel<MLPFeatureExtractor, K>,
    x: &[f64],
    y: &[f64],
    h: f64,
) -> Result<Vec<f64>>
Expand description

Numerical gradient ∂K_DKL/∂θ via central finite differences on the flat parameter buffer. Returns a vector of length kernel.feature_extractor().parameter_count().

The caller must pass an MLPFeatureExtractor (or any extractor that shares the parameters / sync_from_flat contract) — the helper needs to perturb the flat buffer and then push the update back into the layer weights before the next forward pass.