Our modern world is powered by computer chips, from the ones in cars and smartphones to those that help track animals and optimize various industries. As technology continues to advance, there is a constant push to develop faster, more efficient, and innovative chips. To achieve this, some researchers are turning to artificial intelligence (AI) to assist in the chip design process, sometimes taking a step back from human control.
A team of scientists recently shared their process of allowing AI technology to design and test more efficient computer chips. Led by Kaushik Sengupta, an electrical engineer at Princeton University, the research explores the potential of using AI to enhance chip development. Sengupta, who was recently awarded an IEEE fellowship for his work on wireless chip technology, emphasizes that the goal is not to replace human engineers, but rather to augment their productivity. By publishing their findings in the open-access, peer-reviewed journal Nature Communications, Sengupta’s team is making their AI-driven research available to the broader scientific community.
The research team’s use of convolutional neural networks (CNNs), a type of AI, is central to their chip design process. CNNs are typically used in image recognition, but the team has adapted this technology to improve chip designs by generating unconventional and creative layouts. Unlike human engineers, AI can process information in a non-linear fashion, allowing it to propose innovative designs that might never occur to a human mind. Sengupta and his team believe that traditional chip design methods—relying on established templates and form factors—limit innovation. AI, however, can suggest entirely new paradigms in a matter of minutes, giving engineers fresh perspectives to build upon.
The AI system’s designs may appear chaotic and disorganized compared to traditional chip designs, but they offer new possibilities by rethinking the structure and flow of electrical signals. “Classical designs carefully put these circuits and electromagnetic elements together piece by piece,” Sengupta explained, “but by changing those structures, we incorporate new properties. The options are much larger now.”
However, even AI-powered chip designs come with their own set of challenges. AI systems can still make errors, such as “hallucinations” where the AI suggests designs that are not practical or grounded in reality. These shortcomings require human designers to step in and make corrections. Sengupta emphasizes that the purpose of AI in this context is not to replace human designers but to enhance their ability to generate new ideas. “The human mind is best utilized to create or invent new things,” Sengupta said, “and the more mundane, utilitarian work can be offloaded to these tools.”
The goal is for AI-designed chips to complement human ingenuity, offering engineers inspiration and helping them solve complex problems more efficiently. Sengupta envisions a future where AI helps engineers achieve “aha!” moments, enabling breakthroughs that are still understandable and modifiable by human experts.
Because the research is open and transparent, it allows for greater collaboration within the scientific community, offering the potential for more widespread adoption and innovation. While it remains to be seen whether these AI-designed chips will become the backbone of the wireless networks powering our world, their potential to transform chip design and manufacturing is undeniable.
By Impact Lab
