Time-Series Quantum Reservoir Computing with Weak and Projective Measurements

Pere Mujal

Nowadays, classical machine learning algorithms are a necessary tool to process the increasing volume of available big data, including temporal sequences. Prominent time-series processing tasks are speech recognition or stock market and climate forecasting. In this context, a neuromorphic approach known as reservoir computing has become popular. The development of its quantum counterpart, quantum reservoir computing [1], has given rise to theoretical proposals that display superior performances in different platforms, including current noisy intermediate-scale quantum computers. However, a central issue towards the experimental implementation of online time-series processing is the destructive effect of quantum measurements. We tackle this problem by proposing and analyzing different realistic measurement protocols still achieving high performance, enabled by quantum coherence, and efficiency [2].

[1] P. Mujal, R. Martínez-Peña, J. Nokkala, J. García-Beni, G. L. Giorgi, M. C. Soriano, and R. Zambrini, Opportunities in Quantum Reservoir Computing and Extreme Learning Machines, Advanced Quantum Technologies, 2100027 (2021).

[2] P. Mujal, R. Martínez-Peña, G. L. Giorgi, M. C. Soriano, and R. Zambrini, Time-Series Quantum Reservoir Computing with Weak and Projective Measurements, arXiv:2205.06809 (2022).

Pere Mujal got his Ph.D. in Physics from the Universitat de Barcelona in 2019 and he is originally from La Valldan (Berga), Catalonia. Now, in Palma, Mallorca, he is working as a postdoctoral researcher at Institut de Física Interdisciplinària i Sistemes Complexos, IFISC. His research has been focused on the study of interacting ultracold few-boson systems. He is currently interested in the broad field of quantum machine learning and, in particular, he is working on quantum reservoir computing.