[19p-Z03-1~13] 3.10 光量子物理・技術

▼ [19p-Z03-7] N-users optimal resource-sharing with entangled photons

〇Nicolas Chauvet¹、Guillaume Bachelier²、Serge Huant²、Hayato Saigo³、Hirokazu Hori⁴、Ryoichi Horisaki¹、Makoto Naruse¹ (1.Univ. Tokyo、2.Univ. Grenoble Alpes、3.Nagahama Inst. Bioscience and Tech.、4.Univ. Yamanashi)

キーワード：Decision-making, Entangled states, Evolutionary stabiliy

Photonics technology has recently been involved in several studies on artificial intelligence thanks to its ease of use, robustness, and tunability. In particular, single-photon based experiments in decision-making situations have sparked an interest in using quantum states of light directly as a new resource for social situations such as the Competitive Multi-Armed Bandit. In this work, we determine optimal states for the 2-arms N-players situation with polarization-entangled photon pairs, for N up to 5. These states are selected to guarantee fair and optimal outcome among players while minimizing conflict of decision. With these states, selfish actions coincide with socially beneficial consequences, proving the algorithm’s evolutionary stability. Finally, a straight-forward and scalable polarization alignment algorithm guarantees robustness and device-independence.