[PS-1-15] Enhancing Perceptual Artificial Intelligence Systems with a Dynamically Reconfigurable CMOS-Compatible Synaptic Transistor
We propose a reversible artificial synaptic transistor capable of switching between n-type and p-type modes via electrostatic doping. By utilizing program gate (PG) and control gate (CG), we achieve independent control of channel polarity and driving, enabling seamless mode switching. Furthermore, we enhance the device's performance by integrating an inorganic synapse layer into the reconfigurable field-effect transistor (RFET). This high-performance synaptic device exhibits strong responses to pre-synaptic spikes in both modes, demonstrated through excitatory post-synaptic current (EPSC) and paired-pulse facilitation (PPF). These advancements facilitate the development of advanced artificial intelligence (AI) systems with exceptional computational and processing capabilities.