Cornell University researchers have developed a groundbreaking "microwave brain" chip that processes AI tasks and handles wireless communication using microwaves instead of digital logic. This innovative chip slashes power use while boosting speed, paving the way for always-on, cloud-free AI in tiny devices.
The chip's design builds a type of AI framework directly into the hardware, using the natural behavior of microwaves to process incoming data streams. It leverages tunable waveguides to produce a controlled "mush" of frequency behaviors, allowing it to recognize patterns and learn from data. This approach enables the chip to operate at tens of gigahertz, far surpassing most digital chips, while consuming less than 200 milliwatts of power.
The microwave brain chip can decode radio signals, track radar targets, and classify high-speed data streams in real-time with up to 88% accuracy. Its potential applications include embedding advanced computing capabilities directly into consumer devices like smartwatches or cellphones, detecting subtle anomalies in wireless communication across multiple microwave frequency bands for hardware security., and decoding radio signals and tracking radar targets with high accuracy.
This breakthrough could revolutionize the field of AI and wireless communication, enabling the development of more efficient and powerful devices that can process complex tasks in real-time. With its high speed, low power consumption, and versatility, the microwave brain chip has the potential to transform a wide range of industries and applications
