TY  - RPRT
TI  - CLAR: Learning 3D Representations for Robotic Manipulation by Fusing Masked Reconstruction with Multi-Level Contrastive Alignment
AU  - Wenbo Cui
AU  - Chengyang Zhao
AU  - Yuhui Chen
AU  - Haoran Li
AU  - Zhizheng Zhang
AU  - Dongbin Zhao
AU  - He Wang
PY  - 2026
UR  - https://arxiv.org/abs/2507.08262
ID  - 2507.08262
ER  -