Weighted Mean Topological Dimension

The paper explicitly states and proves both targets: mdim_a(X1,T1) ≤ mdim^a_M(X1,d) (Theorem 1.4) and, under SBP, mdim_a(X1,T1)=0 (Theorem 1.5). The proof of Theorem 1.4 builds a Lipschitz partition-of-unity map F(N,·), uses a probabilistic selection (Claim 3.8) and a retraction to a low-dimensional skeleta, to bound the order D of the weighted join and conclude the inequality; see the statement of the main theorems and the proof structure with F(N,·) and Claim 3.8 in the paper . The conclusion D(·) ≤ (D+ε)N+1 appears at the end of the same argument . The SBP part constructs partitions of unity with small boundary orbit capacity (Prop. 4.3) and shows the image lies in a union of low-dimensional affine subspaces, yielding zero weighted mean dimension . The candidate model reaches the same two conclusions via a different route: (i) passing through width dimension and covering-number bounds; (ii) reducing to classical mean dimension zero under SBP and decomposing the weighted join into blocks. The model’s approach is essentially correct but omits a rigorous justification of the weighted analogue of the cover↔width-dimension identity it invokes. Hence, both are correct, with the paper providing a complete constructive proof and the model giving a valid but less detailed alternative.