MoST: Efficient Monarch Sparse Tuning for 3D Representation Learning

Figure 1. Existing 3D parameter-efficient fine-tuning (PEFT) methods rely on additional adapters or prompts, which, while using point cloud priors, introduce inference overhead and lack generalization. Reparameterization-based PEFT methods like LoRA[23], though free of the above issues, overlook point cloud characteristics. MoST combines the best of both worlds by reparameterizing dense update weight matrices with tailored sparse Point Monarch matrices, preserving local geometry, avoiding inference overhead, and remaining generalizable.