Ziqi Zhou

Asynchronous RL training on long-horizon tasks achieves superior performance by leveraging Muon and a novel single-rollout recipe that avoids waiting for complete trajectories.

Standardizing terminal-outcome advantages can significantly boost online learning efficiency in asynchronous reinforcement learning scenarios.

Object-Uni transforms how we understand and generate spatial representations of objects, enabling precise manipulation of their poses in generated images.

Instruction-dense visual jailbreaks can covertly embed harmful instructions in images, bypassing safety measures that protect against direct text generation.