Machenta, one demo and a new task
According to Imperial College London, AI-powered robots are typically trained by humans demonstrating individual tasks many times, often over days. Machenta, a spin-out of the university, says its model lets a robot learn a new task from a single demonstration.
On 27 October, Imperial College London announced that Machenta has raised a $40 million seed round, co-led by Creandum and Union Square Ventures. Imperial College London calls it the largest seed raise by a UK robotics startup. The Next Web reports, citing Business Insider, that the company declined to disclose its valuation.
The work began six years ago in the Robot Learning Lab at Imperial College London. Professor Edward Johns, Machenta's co-founder and CEO, began it with his team there. He later founded the company with his PhD student Vitalis Vosylius and colleagues from the lab.
The model is pre-trained entirely on simulated data. According to Imperial College London, that means future developers may no longer face the practical and privacy constraints of acquiring large volumes of human data. The company says one model has learned tasks such as packing and tidying across different robots, including humanoids. Johns says: "Intelligence is not so much about what you can already do, it's about how quickly you can learn to do something new."
The Next Web also lists limits. Business Insider reported that Machenta has 11 employees, remains focused on research and has no near-term plans to sell its technology. The Next Web adds that more precise tasks and complex movements remain difficult for the model.
What the money pays for: scaling the model and growing the team at a new office and lab in King's Cross, London. Kathi Wilhelm, partner at Creandum, says "very few teams in robotics can point to a body of work like this."