Singapore start-up Sapiens cracks AI training bottleneck, makes waves overseas
The technique is already attracting interest from robotics and medical companies in the United States, China and Singapore.
Singapore-based artificial intelligence company Sapiens Technology has overcome a significant hurdle in the field of robotics by drastically reducing the amount of time needed to train AI models. This achievement, which has garnered attention from global corporations, enables the cutting-edge technology to train AI systems in as little as 30 minutes, a stark contrast to the months or even years traditionally required.
The secret lies in Sapiens' innovative method of training AI on only a fraction of their parameters, rather than the entire system, resulting in quicker training times.
Sapiens is currently in discussions with over 30 robotics and medical firms across the United States, China, and Singapore to pilot and scale its solution. Notable potential customers include China’s TetraBOT Electronic Technology and Shanghai Xinwei Medical Technology, both of which have expressed interest in harnessing the technology's capabilities for their industrial applications.
At the heart of Sapiens' breakthrough is the ability to train AI models using a mere 1 to 3 billion "parameters," as opposed to entire models that can span hundreds of billions, or even trillions, of parameters. This approach not only accelerates the training process but also allows AI systems to adapt to different industries and situations more easily.
By focusing on common elements applicable to various scenarios, such as understanding the environment and self-improvement, Sapiens' technique enables robots to learn and adapt on the fly, even when faced with changes in tasks.
To ensure the safety and integrity of the trained robots, Sapiens has implemented an auditing and monitoring system that allows for the easy interpretation of the AI's logic in natural, readable language. Additionally, a built-in delay between the AI's planning and execution phases enables humans or monitoring systems to halt any unintended or rogue behaviour, adding an extra layer of safety to the technology.
Sapiens' technique has already been tested successfully on commercial robotic arms and off-the-shelf robot ducks performing tasks such as picking up objects. Its potential extends far beyond the realm of robotics, with medical companies like Shanghai Xinwei Medical Technology exploring its applications in forecasting inventory gaps and generating business insights.
By leveraging industry data and a self-learning system, Xinwei aims to adapt AI models to better understand the relationships between consumables and distributors in their procurement pipeline.
Written by urgent.news from Straits Times's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
