These Hong Kong Scientists Have Used Advanced AI To Unlock Human Brain Modeling

In a bid to transcend traditional computing, a team of scientists in Hong Kong is delving into the intricacies of the human brain to inspire the development of advanced artificial intelligence (AI). Led by Li Can, an assistant professor at the University of Hong Kong, the team aims to create AI circuits and systems capable of lifelong learning and versatile task performance, mirroring the human brain’s capabilities.

Li Can emphasizes the fundamental differences between the human brain and conventional computers. While our brains can function despite defects, computers face challenges with malfunctioning transistors. Li sees a crucial distinction in humans’ ability to learn from experience and reason based on vague information, traits that computers, even powerful AI, currently lack.

The team’s exploration focuses on a cutting-edge memory device called a memristor, which mimics the behavior of biological synapses and neurons in the human brain. Li envisions a paradigm shift: “Brain-inspired devices we are working on would result in a technology different from conventional computers. It will also be a promising addition to speeding up graphics processing units [GPUs].” This innovation is poised to enhance the efficiency of GPUs, crucial in training AI systems.

As AI models, exemplified by GPT-4, expand beyond 1 trillion parameters, Li predicts that transferring data will become a bottleneck, emphasizing the increasing importance of memory systems. Here, the memristor proves pivotal, enabling direct computation within memory and eliminating the need for data transfers between memory and processing units. Li envisions future memristor chips offloading AI tasks with remarkable energy efficiency from energy-intensive data centers to devices like smartphones, watches, and even wearable health-monitoring implants.

The potential applications extend to expediting virus genome sequencing and reducing the time needed for sequencing the coronavirus. Li’s research, supported by a HK$5 million (US$640,000) grant from the Croucher Foundation, signifies Hong Kong’s capacity to attract top talent and drive groundbreaking research in brain-inspired computing.