POLYN Technology has published a new white paper that examines in depth the potential of analog computing in artificial intelligence and neural networks to deliver a better energy-saving alternative in dealing with vast volumes of data.
The white paper, titled Analog Computing for Artificial Intelligence: How to Perform MAC Operations Using Ohm’s Law, explores how to perform high-capacity MAC operations – the main calculations used in AI – in a small area while maximizing energy efficiency. It also looks at how Ohm’s Law can be applied to streamline these processes. The white paper is available here.
POLYN is a fabless semiconductor company that supplies application-specific Neuromorphic Analog Signal Processing (NASP) technology and products.
Analog computing – although it was once considered obsolete – is regaining relevance in the context of AI hardware development. As the white paper explains, by leveraging Ohm’s Law for MAC operations, analog systems offer a compact and energy-efficient alternative to digital processors, especially in low-precision applications.
“While technological challenges remain, the rapid pace of innovation in memory elements and novel neuromorphic analog circuitry suggests that analog computing may soon play a significant role in Edge AI,” said Eugene Zetserov, Vice President of Marketing for POLYN Technology.
“Analog AI hardware will likely be applied to mass sensor data pre-processing solutions,” Zetserov added. “In smart sensors and other IoT scenarios, efficiently handling a massive amount of raw data is vital before sending it to the host.”
For more information about POLYN, visit polyn.ai.