Jiajun Liu

Imagined failures in a world model can pinpoint exactly which robot skill components need new demonstrations, boosting success from 13% to 75% with minimal data.

Frontier models don't need retraining to master lifelong embodied tasks—structuring memory handovers and spatial consolidation across fresh reasoning sessions alone drives a 22.6-point surge in multi-task navigation success.

Purely autoregressive VLAs no longer require awkward hybrid diffusion heads to achieve fine-grained robot control: structuring discrete action tokenization as causally annealed flow matching recovers continuous diffusion precision while preserving end-to-end token efficiency.

Current LLMs falter in mobile environments, with performance plummeting under real-world constraints, revealing critical gaps in their planning capabilities.

Agentic control in zero-shot navigation rivals industrial-scale policies, achieving up to 78% success with minimal interfaces and fewer resources.