A research team under China's Tsinghua University has just developed the world's first hybrid artificial intelligence (AI) chip that accommodates both computer-science-based machine-learning algorithms and neuroscience-oriented schemes.
The research paper on the chip, named Tianjic, was published in the Nature journal on Wednesday as its cover story.
The Tianjic chip which features a hybrid architecture, seeks to combine the conventional computing design and neurologically inspired one.
Analysts said the new findings by Chinese scientists could prove crucial to the future development of AI and provide an alternative toward more general forms of AI, showcasing China's rising expertise in the AI chip field.
Researchers generally take one of two paths, via either computer science or neuroscience, to pursue artificial general intelligence, but it's generally considered the best solution to integrate the two approaches and combine their respective advantages, according to an article posted on the website of Tsinghua University.
Tianjic proves this theory by creating a hybrid platform that can not only accommodate machine-learning algorithms based on computer science, but can also implement brain-inspired circuits, according to the research paper's abstract.
The chip's research team has also demonstrated the hybrid chip's processing capabilities with an unmanned bicycle. Equipped with the chip, the bicycle can complete tasks such as receiving voice orders, achieving self-balance, and detecting and tracking objects in front of it so that it can automatically evade obstacles, the abstract said.
According to the research paper, the chip study is expected to "stimulate artificial general intelligence development" by paving the way to more generalized hardware platforms.
