Chameleon-inspired chip adapts to changing data speeds and cuts errors 40-fold

A new semiconductor device from KAIST can change how quickly it responds to incoming data, helping it process signals that vary from fast to slow without relying on extensive software processing. The device, called a programmable dynamic memtransistor (PDM), combines the memory function of a storage device with the computing function of a transistor. Unlike conventional semiconductor devices with fixed response speeds, its temporal response can be programmed to different settings and retained. Researchers used the device to process time-series data, including signals that contained both rapid and slow changes. In these tests, the PDM reduced prediction errors by up to 40 times compared with conventional fixed-response semiconductor devices. The team also built an integrated PDM array with devices configured for different response speeds. This allowed the array to process multiple time-varying signals in parallel while extracting information across different timescales. Chips that adapt to time The ability to change response speed is central to the technology. Data from real-world systems rarely changes at one constant rate. A sensor in a robot, for example, may capture movements that happen in fractions of a second alongside slower changes. Conventional hardware generally has a fixed temporal response after fabrication, meaning software must often compensate when incoming data behaves differently. That can add computational work and consume more power. KAIST’s researchers addressed this by placing two functional layers inside the transistor. One stores and processes incoming charge, while another traps electrons to control how quickly the device returns to its original state. By adjusting the electron-trapping layer, the researchers could tune the device’s current recovery time across an approximately fivefold range. Its characteristic frequency could be adjusted across a range of more than 10 times. The PDM also retains its programmed response characteristics without requiring continuous external power. This gives the device a form of built-in adaptability that conventional fixed-response semiconductor hardware lacks. The researchers tested the technology on data with multiple timescales. When fast and slow patterns were mixed together, the PDM produced substantially lower prediction errors than a fixed-response device. The team also reported that its PDM array reached accuracy comparable to conventional software-based systems while consuming far less energy. Hardware learns signal speeds The approach could be useful in systems that need to process changing data locally, including autonomous vehicles, robots and wearable devices. Instead of sending all incoming data through complex software processing, some of the work could happen directly in the semiconductor hardware. The researchers demonstrated the concept using signals whose characteristics changed over time, including handwriting and object movement at different speeds. The PDM can adjust its response characteristics to match those different input timescales. “This study demonstrates an AI semiconductor whose response characteristics can be programmed to efficiently process data changing at different speeds,” said Chair Professor Shinhyun Choi. “We expect it to become a core technology that improves the performance of AI devices such as autonomous vehicles, robots, and wearables while reducing their power consumption.” The device was developed by a KAIST-led team that included researchers from Samsung Electronics’ Semiconductor R&D Center. Its compatibility with materials used in established semiconductor manufacturing processes could also help with future commercialization. The research was published in the journal Nature Communications. Recommended ArticlesGet the latest in engineering, tech, space & science - delivered daily to your inbox.With over a decade-long career in journalism, Neetika Walter has worked with The Economic Times, ANI, and Hindustan Times, covering politics, business, technology, and the clean energy sector. Passionate about contemporary culture, books, poetry, and storytelling, she brings depth and insight to her writing. When she isn’t chasing stories, she’s likely lost in a book or enjoying the company of her dogs.