Quantum reservoir computing (QRC) exploits the dynamical properties of quantum systems for machine learning tasks. We show that optimal performance in QRC can be achieved without resorting to disordered systems. Typically, QRC uses dynamical systems with all-to-all topology and random coupling to avoid redundancies. However, our work explores the emergence of a chaotic phase resulting from the competition between coupling and interaction terms in the one-dimensional Bose-Hubbard model with homogeneous coupling. Our results challenge established design principles and suggest simpler and more efficient implementations of QRC.
Presential in the seminar room, and Zoom stream at
zoom.us/j/98286706234?pwd=bm1JUFVYcTJkaVl1VU55L0FiWDRIUT09
