Flexible light-stimulated artificial synapse based on detached (In,Ga)N thin film for neuromorphic computing

Area of Science:

• Materials Science
• Neuroscience
• Computer Engineering

Background:

• Flexible artificial synapses are crucial for next-generation neuromorphic computing.
• Developing stable and adaptable synaptic devices is essential for advanced artificial intelligence.

Purpose of the Study:

• To demonstrate a flexible artificial synaptic device.
• To evaluate its performance in learning, memory, and under bending conditions.
• To assess its potential for neural network applications.

Main Methods:

• Fabrication of a flexible synaptic device using lift-off (In,Ga)N thin film.
• Testing of synaptic functions (learning, forgetting, relearning) in both flat and bent states.
• Simulation of a three-layer neural network using experimental synaptic conductance data.

Main Results:

• The device successfully mimicked biological learning, forgetting, and relearning.
• It demonstrated stable excitatory post-synaptic current under bending.
• Simulated neural network achieved a 90.2% recognition rate.
• The device showed transition from short-term to long-term memory under bending.

Conclusions:

• The flexible (In,Ga)N synaptic device exhibits high stability and learning-memory capability.
• This technology holds significant potential for wearable intelligence and flexible neuromorphic systems.