Vision Transformer Autotunes Quantum Dot Hamiltonians Toward Majorana Modes

A physics-informed vision transformer trained unsupervised on synthetic conductance maps can propose Hamiltonian parameter updates that drive quantum dot chains into topological phases hosting Majorana zero modes in a single step. The model generalizes across a broad range of initial detunings and supports iterative refinement from transport measurements.