Kaspersky Laboratory, a leading cybersecurity company, has announced a significant breakthrough in the field of artificial intelligence (AI). The company has developed neuromorphic chips that can simulate the workings of the human brain with remarkable energy efficiency and high data processing capabilities, according to a report by RIA Novosti.
This development addresses the current challenges in AI, which typically relies on graphic processors with tensor cores. The new neuromorphic chips offer a superior alternative, surpassing traditional approaches in several key parameters.
One of the major advantages of these neuromorphic chips is their energy efficiency. For instance, a chip containing 8,000 neurons consumes only 4 milliwatts, compared to the human brain’s consumption of about 20 watts for its 90 billion neurons. This stark contrast becomes even more apparent when compared to GPU-driven data centers, which consume kilowatts and megawatts of energy.
Another crucial advantage is the speed of information processing. When combined with a compatible camera, these neuromorphic chips can process up to an impressive 1,250 frames per second. This makes them ideal for analyzing fast-paced processes, such as the movement of sand grains in an hourglass or tracking raindrops on a window during rainfall.
The potential applications for this new technology are extensive, ranging from industrial thin film spraying to traffic control systems. However, the high cost of neuromorphic cameras currently hinders their widespread adoption by consumers.
Kaspersky Laboratory has already entered the international market with these new chips and is actively seeking partners to collaborate on pilot projects. The company also aims to collaborate with microelectronics experts in order to produce chips with a process size of 28 nanometers or less. Kaspersky hopes to establish similar production capabilities in Russia.