SOURCE-LINKED INTELLIGENCE

3D-MRL: Nested Multimodal 3D Representations via Matryoshka Representation Learning

arXiv · AI, language, vision and robotics · article · Aug 29, 2026 · UTC

Vision-Language Models align point clouds with image and text embeddings, enabling zero-shot recognition, retrieval, and open-vocabulary understanding of 3D shapes. Existing multimodal 3D pre-training methods produce fixed-dimensional embeddings, requiring separate models for different computational budgets. We propose 3D Matryoshka Representation Learning (3D-MRL), a multimodal 3D pre-training framework based on Matryoshka Representation Learning. 3D-MRL learns nested 3D representations by aligning point clouds with frozen CLIP image and text embeddings while applying contrastive supervision

Read original source ↗ Open in workspace

recordType

paper

region

Global