Figure 4 (IMAGE)

Pohang University of Science & Technology (POSTECH)

Figure 4

Caption

Simulation results of MNIST classification using IGZO 2T synaptic devices. High linear and symmetric weight programming characteristics of AI synaptic devices based on oxide semiconductors enable accurate training processes that minimize errors in artificial intelligence systems. In a simulation of random handwriting recognition, these devices achieved an inference accuracy of approximately 98%, which is nearly identical to the accuracy of ideal mathematical calculations.

Credit

POSTECH

Usage Restrictions

None

License

Licensed content

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.