Research map / Vision-language and multimodal / Cross-modal learning
Modality Alignment across Trees on Heterogeneous Hyperbolic Manifolds
Wu Wei, Xiaomeng Fan, Yuwei Wu, Zhi Gao, Pengxiang Li, Yunde Jia, Mehrtash Harandi
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace single visual feature with text-guided hierarchical visual features (Ours-Euc vs ProTeCt) (ablation) | MaPLe · Rare Species TOS 16-shot | HCA: 24.82 -> 51.81 | Improved |
| add hyperbolic constraints to feature trees (Ours-HypV2 vs Ours-Euc) (ablation) | MaPLe · Cifar100 TOS 16-shot | HCA: 68.01 -> 69.33 | Improved |
| add intermediate manifold search (Ours vs Ours-HypV2) (ablation) | MaPLe · Cifar100 TOS 1-shot | HCA: 51.67 -> 53.19 | Improved |
| add intermediate manifold search (Ours vs Ours-HypV2) (ablation) | MaPLe · Rare Species TOS 1-shot | HCA: 20.82 -> 20.94 | No clear effect |
| use independent curvatures vs shared curvature (HypV2 vs HypV1) (ablation) | MaPLe · Cifar100 TOS 1-shot | HCA: 51.86 -> 51.67 | No clear effect |
| add Ours on top of MaPLe vs ProTeCt | MaPLe · ImageNet TOS 1-shot | LA: 68.91 -> 66.33 | Hurt |
| add Ours on top of MaPLe vs ProTeCt | MaPLe · ImageNet TOS 16-shot | HCA: 31.24 -> 43.79 | Improved |
| apply Ours to PromptSRC instead of ProTeCt | PromptSRC · Rare Species TOS 16-shot | HCA: 33.92 -> 50.77 | Improved |
| learn multiple curvatures instead of a single shared curvature (ablation) | MaPLe · training cost | time per batch (s): 74 -> 74.5 | No clear effect |
Closest papers by meaning
- Topology-Aware Representation Alignment for Semi-Supervised Vision-Language Learning (NeurIPS 2026 · similarity 0.89)
- HyperET: Efficient Training in Hyperbolic Space for Multi-modal Large Language Models (NeurIPS 2025 · similarity 0.89)
- Homology Consistency Constrained Efficient Tuning for Vision-Language Models (NeurIPS 2024 · similarity 0.88)
- AlignVLM: Bridging Vision and Language Latent Spaces for Multimodal Document Understanding (NeurIPS 2025 · similarity 0.88)
- DH-Set: Improving Vision-Language Alignment with Diverse and Hybrid Set-Embeddings Learning (CVPR 2025 · similarity 0.88)
- Linguistic-Aware Patch Slimming Framework for Fine-grained Cross-Modal Alignment (CVPR 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
