Author name: Ziqiao Ma, Xueyang Yu, Haoyu Zhen, Yuncong Yang, Joyce Chai, Chuang Gan

arXiv:2604.07350v1 Announce Type: cross
Abstract: Large Chunk Test-Time Training (LaCT) has shown strong performance on long-context 3D reconstruction, but its fully plastic inference-time updates remain vulnerable to catastrophic forgetting and overf…