Exact ReLU realization of tensor-product refinement iterates

We study scalar dyadic refinement operators on R^2 of the form (Vf)(x,y) = sum_{(j,k) in Z^2} c_{j,k} f(2x-j, 2y-k), where only finitely many mask coefficients c_{j,k} are nonzero. Under a fixed support-window hypothesis, we prove that for every compactly supported continuous piecewise linear seed g:R^2->R, the iterates V^n g admit exact ReLU realizations of fixed width and depth O(n). This gives a first genuinely two-dimensional extension of the exact realization theory for refinement cascades. Using the one-dimensional exact loop-controller framework, the proof transports the tensor-product residual dynamics exactly on the product of two polygonal loops and reduces the remaining seam ambiguity to a final readout and selector step. The matrix cascade is then handled by a fixed-depth recursive block, and general compactly supported continuous piecewise linear seeds are reduced to a finite decomposition together with exact clamped gluing on the support window. This identifies the tensor-product dyadic case as a natural first multivariate instance of the loop-controller method for refinement iterates.

Paper

References (8)

05Composing a depth-O ( n ) network with a fixed affine input translation or with a fixed CPwL output map does not change the fact that the depth remains O ( n )
06i) Every fixed compactly supported CPwL function on R 2 has an exact finite ReLU realization
07(ii) χ 0 ,n ( E ( t )) = χ [0 , 1 / 2) ( t ) and χ 1 ,n ( E ( t )) = χ [1 / 2 , 1] ( t ) for every t ∈ [0 , 1]J
08(ii) Finite sums of fixed-width depth-O ( n ) networks can be realized by enlarging the width by a constant factor while preserving the depth bound O ( n )

Similar papers

© 2026 NYSGPT2525 LLC