Curator's Take
AI Commentary
This article shows how artificial‑intelligence search can uncover new heralded linear‑optical circuits that generate path‑entangled photon states with probabilities far exceeding existing designs, often by exponential factors. By packaging these solutions into a scalable family that subsumes earlier constructions, the work bridges a long‑standing gap between theoretical proposals and experimentally viable multiphoton entanglement, opening a practical route to larger photonic resources for quantum communication and sensing. The results also illustrate how automated discovery can yield transferable design principles, though implementation will still hinge on low‑loss components and high‑efficiency detectors.
— Mark Eatherly
Summary
Entangled states of light lie at the heart of photonic quantum technologies, from distributed quantum communication to quantum-enhanced measurement and information processing. Their practical generation, however, remains constrained by the weak interactions between photons, which make the deterministic assembly of large multiphoton entangled states a central challenge in quantum optics. In this work, we use AI techniques to discover heralded linear-optical schemes for path-entangled states and show that the resulting solutions can be elevated from individual circuits to a new scalable family. This family contains previously known constructions as special cases while generally providing exponential and super-exponential improvements over those, and its extension to broader classes of target states shows how automated discovery can reveal transferable physical understanding. By presenting compact experimental proposals for large path-entangled states, our results provide both a theoretical advance in photonic heralding and a route towards a substantial leap in experimentally accessible multiphoton entanglement.