Scientists in Beijing have developed the world’s first fully optical AI chip, known as Taichi-II. This innovative chip represents a significant advancement in efficiency and performance, surpassing even the renowned NVIDIA H100 GPU in energy efficiency. The Taichi-II chip is a substantial improvement from its predecessor, the Taichi chip, which had already set impressive records in energy efficiency earlier this year.
The research team, led by Professors Fang Lu and Dai Qionghai from Tsinghua University, unveiled their findings on Wednesday. The study highlights Taichi-II’s capability to transform AI training and modeling by leveraging optical processes, which are more efficient than traditional electronic methods. In practical terms, the Taichi-II chip has demonstrated remarkable advancements, expediting the training of optical networks containing millions of parameters by an order of magnitude and improving the accuracy of classification tasks by 40 percent.
Taichi-II advances optical AI training
A key innovation of the Taichi-II chip is its use of a novel approach called fully forward mode (FFM) learning. This technique allows for a computer-intensive training process to be conducted directly on the optical chip, enabling parallel processing of machine learning tasks.
Xue Zhiwei, lead author of the study and a doctoral student, emphasized that this architecture supports high-precision training and is well-suited for large-scale network training. Professor Fang Lu stated, “Our research envisions a future where these chips form the foundation of optical computing power for AI model construction.”
The timing of Taichi-II’s debut is particularly notable as the US has imposed restrictions on China’s access to advanced GPUs for training. The Taichi-II chip offers a viable alternative that could help mitigate these limitations.
This development marks a significant milestone for optical computing, moving it from theoretical concepts to practical, large-scale applications. The convergence of AI with physics could well be the key to unlocking the next phase of technological evolution, balancing efficiency with environmental sustainability.
