Covariant claims its AI model will let robots learn like humans

12 Mar 2024

Image: © Nay/Stock.adobe.com

The company claims its RFM-1 model combines general internet data with data on physical interactions to help robots make decisions ‘on the fly’.

AI robotics company Covariant has revealed its new foundation model, which it claims can give robots the “human-like ability” to reason.

Covariant said its AI model – RFM-1 – utilises the power of generative AI to give robots a better understanding of language and the physical world. The company aims to tackle the biggest issues in modern robotics: reliability and flexibility.

Robots are already used in various factories and manufacturing sites for specific, mundane tasks, but the concept of more general-purpose robots is a much steeper hill to climb. Companies like Google have shown middling success in this field, while various start-ups claim to be working on the next big robotics breakthrough.

Covariant noted that there have been recent advances in foundation AI models and that they are pre-trained on “millions of tasks, represented by trillions of words from the internet”.

“However, partly due to the limitation of their training data, existing models still struggle with grasping the true physical laws of reality, and achieving the accuracy, precision and reliability required for robots’ effective and autonomous real-world interaction,” the company said.

Covariant said RFM-1 is trained on both general internet data and data that is “rich in physical real-world interactions”. The company claims this gives robots the sophisticated ability to reason and make decisions “on the fly”. The model includes 8bn parameters trained on text, images, videos, robot actions, and “a range of numerical sensor readings”.

“We built a highly scalable data collection system which has collected tens of millions of trajectories by deploying a large fleet of warehouse automation robots to dozens of customers around the world,” said Covariant co-founder and CEO Peter Chen.

Covariant’s initial focus is on warehouse operations but said it could be used for broader applications such as hospitals, factories, shops and in homes.

Meanwhile, tech giants appear to be taking a greater interest in the robotics sector, as OpenAI, Microsoft and others recently invested in Figure AI, a start-up focused on developing autonomous, general-purpose humanoid robots.

Find out how emerging tech trends are transforming tomorrow with our new podcast, Future Human: The Series. Listen now on Spotify, on Apple or wherever you get your podcasts.

Leigh Mc Gowran is a journalist with Silicon Republic

editorial@siliconrepublic.com