Advertisement
Artificial intelligence
TechTech Trends

SenseTime pushes open-source AI model to provide a ‘smart brain’ for every robot

The Chinese AI pioneer also releases first robot dogs under new venture Ace Robotics

Reading Time:2 minutes
Why you can trust SCMP
The first robot dog products unveiled on Thursday by Ace Robotics, a new venture of SenseTime. Photo: Handout
Vincent Chow
Chinese artificial intelligence pioneer SenseTime has open-sourced a “world model” and kicked off its foray into the burgeoning field of embodied intelligence by releasing its first robotic dogs.
The Hong Kong-listed AI firm’s new venture, Ace Robotics, on Thursday released Kairos 3.0, an open-source model that simulates the physical world, and a developer platform called “KaiWu” that supports embodied intelligence applications.
“Ace Robotics will promote the large-scale commercialisation of embodied intelligence and build an independent and controllable development path for China to lead the world in this technology,” said SenseTime co-founder and executive director Wang Xiaogang, who also serves as chairman at Ace Robotics.
Advertisement

According to the SenseTime-backed venture, Kairos 3.0 would give every robot a “smart brain”, as the open-source world model “understands the physical laws of the real world”.

The product announcement reflected the growing interest of prominent AI companies – including Tencent Holdings, Google DeepMind, Elon Musk’s xAI and SenseTime – in the field of “spatial intelligence”, which would provide virtual environments to train AI-powered robots.
Advertisement

SenseTime’s wealth of real-world data would give it an advantage in developing useful robots through powerful world models, Wang said in a group media briefing.

SenseTime co-founder Wang Xiaogang, who also serves as chairman at Ace Robotics. Photo: SenseTime
SenseTime co-founder Wang Xiaogang, who also serves as chairman at Ace Robotics. Photo: SenseTime
Advertisement
Select Voice
Choose your listening speed
Get through articles 2x faster
1.25x
250 WPM
Slow
Average
Fast
1.25x