Engineers develop a high-performance and high-reliability artificial synaptic semiconductor device

The neuromorphic computer system technology that mimics the human brain must overcome the excessive power consumption limitation that characterizes the existing von Neumann computing method. A high-performance analog artificial synapse device capable of expressing synapse connection strength is required to implement a semiconductor device using a brain information transmission method. This method uses signals transmitted between neurons when a neuron generates a spike signal.

In conventional variable resistance memory devices widely used as artificial synapses, as the filament grows with varying resistance, the electric field increases, causing a feedback phenomenon, resulting in rapid growth of the filament. Therefore, it is a challenge to implement plasticity while maintaining analogous (gradual) resistance variation with respect to filament type.

The Korea Institute of Science and Technology, led by Dr. YeonJoo Jeong’s team at the Center for Neuromorphic Engineering are addressing the limitations of analog synaptic properties, plasticity, and information retention that are chronic barriers to memristors, neuromorphic semiconductor devices. He announced the development of an artificial synaptic semiconductor device that enables highly reliable neuromorphic computing.

The KIST research team refined the redox properties of active electrode ions to solve small synaptic plasticity problems affecting the performance of existing neuromorphic semiconductor devices. In addition, transition metals have been doped and used in the synaptic device to control the reduction probability of active electrode ions. Engineers discovered that the high reduction probability of ions is a critical variable in designing high-performance artificial synaptic devices.

Therefore, a titanium transition metal having a high ion reduction probability was introduced into an existing artificial synaptic device by the research team. This preserves the analogous properties of the synapse and the plasticity of the device at the biological brain synapse, approximately five times the difference between high and low resistances. In addition, they developed a high-performance neuromorphic semiconductor that is about 50 times more efficient.

Due to the high alloying response exhibited by the doped titanium transition metal, the information retention increased up to 63 times compared to the existing artificial synaptic device. In addition, brain functions, including long-term potentiation and long-term depression, could be more accurately simulated.

The team implemented an artificial neural network learning pattern using the developed artificial synaptic device and tried artificial intelligence image recognition learning. The error rate was reduced by more than 60% compared to the existing artificial synaptic device; In addition, the accuracy of handwriting image pattern recognition (MNIST) increased by more than 69%. The research team confirmed the feasibility of a high-performance neuromorphic computing system through this improved artificial synaptic device.

dr Jeong of KIST said, “This study drastically improved synaptic range of motion and information retention, which were the major technical barriers of existing synaptic mimics. In the developed artificial synapse device, the analog operation range of the device to express the various connections of the synapses The strengths have been maximized, so that the performance of brain simulation-based artificial intelligence computing is enhanced.

“In the follow-up research, we will manufacture a neuromorphic semiconductor chip based on the developed artificial synapse device to realize a high-performance artificial intelligence system, thereby further enhancing the competitiveness in the field of domestic artificial intelligence systems and semiconductors.”

The study was published in nature communication.

Provided by the National Research Council of Science & Technology