Oshio Reon, Kuwahara Takumi, Aoki Takeru, Kimura Mutsumi, Nakashima Yasuhiko
Nara Institute of Science and Technology (NAIST), Ikoma, Japan.
Ryukoku University, Otsu, Japan.
Sci Rep. 2025 Jan 31;15(1):3954. doi: 10.1038/s41598-025-87924-6.
Artificial intelligences are indispensable social infrastructures, neural networks are embodiment methodologies, and neuromorphic systems are promising solutions for compact size and low energy. Memristors were first prepared for the synapse devices but incur energy consumption, and memcapacitors were next prepared but have small dynamic ranges of capacitance. In this research, we have developed a neuromorphic system using capacitor synapses. Here, multiple capacitors have binary-weighted capacitances and are controlled to be connected to intermediate signals. They are discharged through transistors, and when they fall below the threshold voltage, the output signals are inverted. After all, electric charges in the multiple capacitances are summed and measured by the inverting intervals, which is the same as multiply-accumulate operation. A large-scale integration chip is actually fabricated. The working is confirmed by MNIST, and the circuit-aware rounding improves the accuracy to 96%, indicating a sufficient possibility for practical applications, and the energy efficiency is 163 GOPS/W even by the 180 nm technology, indicating a great potential for low energy consumption.
人工智能是不可或缺的社会基础设施，神经网络是具身化方法，而神经形态系统对于紧凑尺寸和低能耗来说是很有前景的解决方案。忆阻器最初是为突触器件制备的，但会产生能量消耗，接下来制备的忆容器电容动态范围小。在本研究中，我们开发了一种使用电容突触的神经形态系统。这里，多个电容器具有二进制加权电容，并被控制连接到中间信号。它们通过晶体管放电，当电压降至阈值电压以下时，输出信号反转。最终，多个电容中的电荷通过反转间隔进行求和与测量，这与乘法累加操作相同。实际制造了一个大规模集成芯片。通过MNIST验证了其工作情况，电路感知舍入将准确率提高到了96%，表明其具有足够的实际应用可能性，并且即使采用180纳米技术，能量效率也达到了163 GOPS/W，显示出低能耗的巨大潜力。
