GEN-1.5: Embodied Foundation Models are One-Shot Learners

Worauf sich diese Seite beruft, wörtlich, abgerufen am 04.09.2026:

• GEN-1.5 is a large multimodal model that processes video input (30 seconds of memory, alongside other sensor, language, and proprioceptive inputs) and produces 100 Hz action trajectories.bestätigt 10.11.2026
• it can learn a new task in seconds, from a single example, without gradient updates or fine-tuningbestätigt 10.11.2026
• this is the first model we know for which one-shot and few-shot learning of physical skills have emerged at scalebestätigt 10.11.2026
• Experiments across 10 diverse tasks show 59% (±10% std. dev.) average success with one-shot in-context prompting, straight from the pretrained model.bestätigt 10.11.2026
• Although the tasks are simple and short-horizon, and the success rates are modestbestätigt 10.11.2026
• With few-shot learning, performance rises to 83% (±9% std. dev.) via 10 gradient steps on 5 minutes of data per task (~50 demonstrations).bestätigt 10.11.2026
• GEN-1.5 pretraining contains no simulation data, neither rendered video nor simulated dynamicsbestätigt 10.11.2026