Physically Driven Curve Modulation of Gaussian TransistorBased on DNTT/IGZO Heterostructure Anti-ambipolar Operation
발표자
()
초록
내용
Large language model (LLM) has rapidly gained interest with the remarkable emergence of artificial intelligences (AIs) driven by Deep neural networks (DNNs), which are now widely used in various industrial fields. However, the credibility of conventional LLMs has been a prominent issue, since they tend to simply present the results without indicating probability. To address this issue, the Gaussian mixture model (GMM) was introduced, where they use individual Gaussian distributions of multiple inputs and weights to generate the results in the form of distribution with digitized certainty. Hence, we propose a p-n heterostructure anti-ambipolar transistors (AAT) based on oxide IGZO and organic DNTT semiconductors, which generates the peak-shape transfer curve with a single device. The effects of physical parameters such as junction overlap area and p/n length ratio were investigated to demonstrate how structural parameters modulate the transistor operation.