Researchers at the University of Florida have developed a groundbreaking photonic chip that utilizes laser light instead of electricity to perform convolution operations, a core function in artificial intelligence (AI). This innovation significantly enhances energy efficiency, achieving up to 100 times the performance of traditional electronic chips.
The chip integrates microscopic Fresnel lenses directly onto a silicon wafer, enabling laser light to execute convolutions with minimal energy consumption. In tests, the prototype accurately classified handwritten digits with approximately 98% accuracy, matching the performance of conventional chips.
This development marks a significant step toward sustainable AI computing, addressing the growing energy demands associated with large-scale AI models. The research team, led by Professor Volker J. Sorger, envisions this technology as a catalyst for future AI systems that are both powerful and energy-efficient.
The University of Florida’s advancement in photonic computing underscores the potential of optical technologies in transforming AI hardware, paving the way for more sustainable and scalable AI applications.
Read more on Tech Gist Africa:
Ghana to Become Africa’s First AI-Powered Agricultural Hub with $100M Japanese Investment
Thoma Bravo to Acquire Verint in $2 Billion Deal, Merging AI-Driven Customer Experience Platforms
