Research On The Properties Of Quantum Synchronization Based On Machine Learning

The origin of the synchronization theory comes from the simple pendulum synchronization phenomenon discovered by C.Huygens in the 17th century.In the hundreds of years after this,the synchronization phenomenon has been developed and improved by a large number of scholars,and has shown its application value in many fields.In recent years,researchers have observed similar synchrony on a microscopic scale.Relying on the unique properties of quantum mechanics,the classical synchronization theory is difficult to accurately describe and analyze the synchronization phenomenon of quantum system.Therefore,quantum synchronization theory has become a new research hotspot.Quantum synchronization theory is an interdisciplinary subject of quantum mechanics,informatics and cybernetics.Its main research objectives are to give the definition and measurement of quantum synchronization phenomenon,and to design the scheme of quantum system to achieve synchronization,etc.However,it is difficult to observe the quantum system directly,and quantum synchronization detection,as an important technology of quantum detection,has been used in the detection or measurement of some quantum systems.For a good quantum detection,the detection scheme is very important.The properties of the system can be inferred from the scheme of interaction between external detectors and quantum synchronization.Quantum synchronization occurs spontaneously when a qubit in contact with the environment interacts with a detection qubit.Base on this scheme,we build a detection model to describe the quantum bit system in the cavity field,and reveal the influence of the environment(i.e.the cavity field)on the quantum synchronization and the interaction between the environment,the qubit system and the detection device.By adjusting the frequency of the probe,in-phase synchronization,anti-phase synchronization and desynchronization can be achieved.On this basis,a variety of different synchronization laws are realized by changing some parameters which affect the quantum synchronization,such as the ohmic dissipation index,the interaction between the cavity field and the system₃ and the temperature of the external environment.With the rapid development of machine learning,it has been able to serve the research of theoretical physics well.A variety of algorithms,such as linear regression,artificial neural network and random forest,make it easier for us to study these problems.In the thesis,we use machine learning algorithm to deal with these hyperparameter problems based on these quantum synchronization theories.The classification and regression of three parameters representing quantum synchronization properties are carried out by machine learning method respectively.The classification error MSE and prediction error NME are analyzed,so that machine learning can better serve the quantum model.The establishment of a complete machine learning model makes it easier to judge the state of quantum synchronization in the future,which provides help for the application of the combination of machine learning and quantum synchronization in the future.