Noticias

learning

Helm.ai Signs $70 Million Physical-AI Model Contracts

Helm.ai says it has signed $70 million in commercial contracts over 12 months for its physical-AI foundation models, with global automakers, Tier 1 suppliers and industrial automation companies. Its unsupervised-learning “Deep Teaching” approach separates environmental perception from action. The company says its software is used in autonomous-vehicle programs and industrial perception, with development extending to robotics, mining, construction and delivery drones.

Traducción pendiente · Resumen en inglés

Robot NewspaperOct 23, 2026Detalles

learning

Inha University SPARO Wins IROS 2026 Workshop Award and Places Second in IMU Challenge

Inha University’s SPARO lab received the IROS 2026 AI Meets Autonomy workshop Best Paper Award for a vision-language-model method that detects LiDAR–camera misalignment across different environments without additional training. Team SPARO also placed second among 131 teams in the IMU Odometry Challenge, estimating motion across vehicles, quadruped robots, drones and handheld devices with one learned model. The lab additionally presented radar localization, collaborative SLAM and underwater depth-completion research.

Traducción pendiente · Resumen en inglés

Robot NewspaperOct 23, 2026Detalles

learning

Dream Labs raises $90M and launches Dream API

Dream Labs announced $90 million in seed funding and launched Dream API, which it says converts a robot camera feed and plain-English instructions into motor commands. Its Dream-1 model is trained on video of people performing everyday tasks and is claimed to transfer across robot hardware without robot-specific data. The company says it currently handles tabletop tasks, including picking and placing objects and folding laundry.

Traducción pendiente · Resumen en inglés

The Robot ReportOct 22, 2026Detalles

learning

Video Friday: Robot Decommissioning Takes a Fun Turn

The roundup highlights Workhorse, a system that learns contact-rich whole-body manipulation from human demonstrations without requiring a robot during demonstration. On a real Unitree G1, it sorts boxes using its hands and a kick, catches a thrown box, and topples and climbs a suitcase. The article also features MIT’s SANDO navigation system and a soft underwater robot with three swimming modes.

Traducción pendiente · Resumen en inglés

IEEE SpectrumOct 9, 2026Detalles

Verified published records only; coverage is not a census.