Shuran Song

This research focuses on optimizing language model training and predicting their real-world performance.

Robots are learning to adapt and perform tasks across various fields.

A new method for reconstructing complex objects using visual input and coding techniques.

A new system helps robots learn tasks using audio from real-life demonstrations.

Sharing experiences to improve future robotic development.

GET-Zero enhances robots' adaptability across varied designs without extensive retraining.

Generative models create diverse training data, enhancing robot adaptability.