Dynamical phase transitions in quantum Reservoir Computing

Aug. 23, 2021

Information can be processed by physical, chemical and biological systems with a major example being our brain. In this context, quantum reservoir computing is a promising approach in the fields of complex systems and artificial neural networks characterized by a fast and easy training and with applications as time series processing and forecasting. Complex systems can store memory and can be used to non-linearly process input information, but understanding what is essential for a good quantum reservoir computer is key in order to achieve experimental implementations.

A team of researchers of the Institute of Cross-disciplinary Physics and Complex Systems (IFISC, UIB-CSIC) has published a paper in Physical Review Letters in which they studied what dynamical natural properties of a qubit quantum network determine its ability to act as a reservoir computer.

The paper shows the importance of tuning the reservoir at the onset of thermalization, which can be easily achieved by controlling the strength of the magnetic field. To do this, the researchers identified different response regimes offered by the model, being able to make dynamic phase transitions from one to another by changes of the degree disorder and transverse field strength. Indeed, a complex qubit network can be considered as a many-body system that can thermalize or display many-body localization. Bridging the context of quantum dynamical phases and neuromorphic computing, how these phenomena play a role in reservoir computing applications and implementations as in ion experiments, where these phenomena have been recently reported. An extensive analysis shows that thermalization is an essential feature for a quantum reservoir computer. Researchers also found that being at the edge of the dynamical transition can be beneficial for the model, while localization hinders information processing and spreading within the physical system.

Uncovering the underlying physical mechanisms behind optical information processing capabilities of spin networks is essential for future experimental implementations and provides a new perspective on dynamical phases. 

Martínez Peña, Rodrigo; Giorgi, Gian Luca; Nokkala, Johannes; Soriano, Miguel C.; Zambrini, Roberta. Physical Review Letters 127, 100502 (1-7). https://doi.org/10.1103/PhysRevLett.127.100502


Related Research projects


Machine learning with quantum reservoir computing

P.I.: Roberta Zambrini
Project funded by the government of the Balearic Islands with the goal of extending Reservoir Computing into the quantum domain. The project is mainly theoretical/numerical, but also deals with the identification of ...


Quantum machine learning using reservoir computing

P.I.: Miguel C. Soriano, Roberta Zambrini
The QuaResC project engages in a new collaboration UIB and CSIC researchers at IFISC with the objective to address an interdisciplinary topic between artificial intelligence and quantum physics: quantum machine learning using ...

This web uses cookies for data collection with a statistical purpose. If you continue browsing, it means acceptance of the installation of the same.

More info I agree