Behavior control of coherent-type neural networks by carrier-frequency modulation

Area of Science:

• Artificial Intelligence
• Optoelectronics
• Quantum Computing

Background:

• Traditional artificial neural networks (ANNs) often lack dynamic control mechanisms.
• Carrier-frequency modulation offers a novel approach to information encoding and processing.
• Self-homodyne circuits provide a framework for coherent signal manipulation.

Purpose of the Study:

• To propose and investigate coherent-type artificial neural networks controlled by carrier-frequency modulation.
• To explore the learning capabilities of these networks using frequency as a parameter.
• To demonstrate the potential for advanced applications in signal processing and computing.

Main Methods:

• Development of artificial neural networks operating based on coherent principles.
• Implementation of carrier-frequency modulation to control network behavior.
• Learning process achieved by adjusting neural connection delay times and conductance.
• Formation of a self-homodyne circuit for the network system.

Main Results:

• Successful control of network behavior demonstrated through carrier-frequency modulation.
• The network effectively learns teacher signals associated with information-carrier frequencies.
• Experimental validation of the proposed coherent-type artificial neural network architecture.

Conclusions:

• Carrier-frequency modulation is a viable method for controlling artificial neural network behavior.
• The developed network architecture shows promise for signal processing applications.
• Future potential extends to frequency-multiplexed optical neural computing and quantum neural devices.