Answers are grounded in this paper and cite their sources.
Theoretical modeling reveals unified scaling beta functions in Anderson localization models, indicating how disjoint asymptotic regimes can be smoothly reconciled.
We formulate a finite-gluing construction of scaling β functions for Anderson localization from local pieces of information that are reliable in different regimes. The construction does not require a single effective theory or a single resummation formula to remain valid over the full conductance range. The insulating asymptote β~ln g, the diffusive metallic asymptote β→ d-2, and a critical local relation around a fixed point are treated as separate inputs and are joined by positive statistical-kernel weights. The critical input may be supplied by a nonlinear sigma model, Borel--Pad\'e resummation, numerical finite-size scaling, or any other local description. Fermi-, Bose-, and para-statistical forms are used only as generators of smooth gluing kernels. As a proof of concept we revisit both the orthogonal and symplectic Wigner--Dyson classes, with the symplectic class providing the main zero-set example. The orthogonal--symplectic Borel--Pad\'e relation is used only to provide nonlinear critical-shape information inside a finite window, while the critical slope 1/ν(d) may be supplied independently. Using the two- and three-dimensional symplectic critical exponents together with the high-dimensional condition ν→1/2, a minimal square-root interpolation of 1/ν(d) yields the illustrative lower-critical-dimension estimate dₗ(sq)=1.705044…. This number is a model value rather than a four-digit physical determination: the present uncertainty is dominated by the choice of dimension interpolation and gluing family. In the corresponding minimal zero-set structure, the unstable critical and stable fixed points coexist for dₗ<d<2, merge at d=dₗ, and the stable branch moves to infinite conductance as d→2⁻. The resulting glued functions recover both insulating and metallic asymptotes and allow the zero set β(d,g)=0 to be followed continuously in dimension. Raw, unnormalized β functions are shown alongside the normalized high-dimensional display, and the new curves are compared directly with the 2017 Ueoka--Slevin Borel--Pad\'e functions. Different admissible kernels share the same imposed local and asymptotic data but can generate different intermediate fixed-point topologies. This nonuniqueness is interpreted not as a defect of the construction, but as an explicit diagnostic of which additional numerical or experimental data are required to select the physical β function. A normalized representation β=β/(d-1) is also introduced for high-dimensional visualization. The method is not restricted to the Wigner--Dyson classes. We survey the chiral and Bogoliubov--de Gennes sectors and show why their local inputs must be class specific: chiral classes possess the Gade--Wegner marginal conductance at the symmetry point, classes D and DIII show weak antilocalization and can support a two-dimensional thermal metal, while topological and Wess--Zumino terms can require additional running variables. These features do not invalidate finite gluing; rather, they determine which outer and critical local relations must be supplied.
No takes yet. Share an insight, caveat, or question.
Yoshiki Ueoka (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: