Humanoid
Innodata Inc. has opened a new laboratory dedicated to capturing motion data for training the next generation of human-like robots, the company announced today. The facility will generate data for training AI that drives humanoid robots and will also independently validate the performance data that robots produce internally.
The Ridgefield Park, N.J.-based data engineering company, founded in 1988, positions itself as providing the high-quality data, evaluation frameworks, and human expertise needed to build AI systems that builders and adopters can trust at scale. According to Innodata, physical AI faces a fundamental data problem: while large language models benefited from the vast amount of data available on the Internet, robots have no equivalent corpus of physical-world data.
Physical AI has to earn its tokens one interaction at a time, and they have to be deliberate.
Franklin Tanner, Innodata’s vice president of robotics and physical AI, explained that physical AI must “earn its tokens one interaction at a time, and they have to be deliberate.” The new research and development facility addresses what Innodata describes as an urgent need for high-quality training data for humanoids, industrial robots, and other forms of physical AI. The company said that while other data providers infer 3D motion by analyzing 2D video, its approach captures 3D data directly from bodies themselves, whether human or otherwise.
CEO Rahul Singhal framed the lab as a response to a bottleneck facing robotics teams. “Physical AI is growing faster than any other segment in AI, but every robotics team hits the same wall: There isn’t enough real-world interaction data, and what exists is expensive and slow to produce,” he stated. “This facility removes that wall.” Singhal said Innodata now offers physical AI companies a complete end-to-end capability — from data collection through model evaluation — that compresses development cycles and gets more capable, safer robots into the world sooner.
Source: The Robot Report