[16a-A33-1~11] FS.1 Focused Session "AI Electronics"

11:00 AM - 11:15 AM JST(2:00 AM - 2:15 AM UTC)Presentation by Applicant for JSAP Young Scientists Presentation Award

[16a-A33-8] Circuit evaluation for quantum reservoir via similarity between output density matrix and maximally mixed state

〇Hideaki Oba¹, Ryuji Sakai¹, Kazuki Uematsu¹, Hideyuki Nakagawa¹, Tomoyuki Takeguchi¹, Yutaro Iiyama², Lento Nagano², Ryu Sawada², Junichi Tanaka², Koji Terashi² (1.R&D Center, Toshiba, 2.ICEPP Tokyo Univ.)

Keywords:

quantum reservoir computing

A quantum reservoir is less affected by real-device noise and thus is attracting attention as an algorithm for NISQ devices since it has no learning parameters in the circuit. However, due to numerous candidates of the quantum reservoir circuits, it is desired to judge suitable circuits. In this presentation, we propose a method to judge whether a reservoir circuit is suitable or not based on the similarity between the output density matrix and the maximally mixed state, leveraging the fact that the performance significantly deteriorates when the quantum reservoir outputs the maximally mixed state. We also discuss the effects of the calculation methods for the density matrix and of the task setting to the judgement of the circuits.