Tianyu Zhou

Photonic reservoir architecture search no longer requires brute-force circuit simulation: an exact polynomial bridge predicts finite-shot risk directly from task statistics to machine precision, while formally proving that standard displacement encodings can never compute nonlinear functions regardless of circuit topology.

Recycling just four transformer layers through twelve recurrent iterations matches full-depth baselines on core linguistic benchmarks, exposing precisely where compute can—and cannot—substitute for raw parameter capacity.