Hyperbolic Contrastive Learning for Hierarchical 3D Point Cloud Embedding
This work addresses the challenge of integrating 3D point clouds with text and images for hierarchical embedding, which is incremental as it builds on existing hyperbolic methods for multi-modal pre-training.
The paper tackles the problem of modeling hierarchical structures in multi-modal data by extending hyperbolic contrastive learning to include 3D point clouds, resulting in a 3D point cloud encoder that significantly improves performance on downstream tasks.
Hyperbolic spaces allow for more efficient modeling of complex, hierarchical structures, which is particularly beneficial in tasks involving multi-modal data. Although hyperbolic geometries have been proven effective for language-image pre-training, their capabilities to unify language, image, and 3D Point Cloud modalities are under-explored. We extend the 3D Point Cloud modality in hyperbolic multi-modal contrastive pre-training. Additionally, we explore the entailment, modality gap, and alignment regularizers for learning hierarchical 3D embeddings and facilitating the transfer of knowledge from both Text and Image modalities. These regularizers enable the learning of intra-modal hierarchy within each modality and inter-modal hierarchy across text, 2D images, and 3D Point Clouds. Experimental results demonstrate that our proposed training strategy yields an outstanding 3D Point Cloud encoder, and the obtained 3D Point Cloud hierarchical embeddings significantly improve performance on various downstream tasks.