Sony AI's Project Ace Masters Real-World Robotics, Rivals Elite Humans
Sony AI published Project Ace, an autonomous robotics system trained via reinforcement learning in simulated then real environments. It competes with elite players in table tennis, achieving rally durations over 100 strokes. The system uses multimodal sensors and adaptive policy networks for physical interaction.
Real-world RL transfer from sim-to-real is now viable for dexterous tasks. Update your robotics workflow: emphasize domain randomization in sims to bridge gaps. This expands AI from digital to physical domains, enabling practical automation.
Sony AI's robotics division deployed Ace on a custom robotic arm. It sustained rallies averaging 120 strokes against professional players, surpassing prior benchmarks by 3x in duration and adaptability.
Step 1: Install Isaac Gym via NVIDIA Hub (developer.nvidia.com/isaac-gym). Step 2: Clone Sony's open-source Ace repo and train policy with PPO algorithm on table tennis sim for 10M steps. Step 3: Transfer to real FR3 robot arm via sim-to-real adaptation; expect 80+ stroke rallies. URL: https://github.com/sonyairesearch/ace-robotics.