Research map / Vision-language and multimodal / Cross-modal learning
Prototype-based Aleatoric Uncertainty Quantification for Cross-modal Retrieval
Hao Li, Jingkuan Song, Lianli Gao, Xiaosu Zhu, Heng Tao Shen
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 |
|---|---|---|---|
| add uncertainty loss (ablation) | PAU · MSR-VTT t2v | R@1: 45.6 -> 47.0 | Improved |
| add uncertainty loss (ablation) | PAU · MSR-VTT v2t | R@1: 45.2 -> 47.4 | Improved |
| add diversity loss on top of uncertainty loss (ablation) | PAU · MSR-VTT t2v | R@1: 47.0 -> 48.4 | Improved |
| add re-ranking with uncertainty vectors to the baseline (ablation) | CLIP baseline · MSR-VTT v2t | R@1: 45.2 -> 46.0 | Improved |
| add re-ranking to full loss model (ablation) | PAU · MSR-VTT t2v | R@1: 48.4 -> 48.5 | No clear effect |
| exponential vs ReLU evidence function (ablation) | PAU · MSR-VTT t2v | R@1: 46.6 -> 48.5 | Improved |
| exponential vs Softplus evidence function (ablation) | PAU · MSR-VTT t2v | R@1: 47.6 -> 48.5 | Improved |
| increase prototypes K from 1 to 8 (ablation) | PAU · MSR-VTT t2v | R@1: 47.2 -> 48.5 | Improved |
| increase prototypes K from 8 to 16 (ablation) | PAU · MSR-VTT v2t | R@1: 48.3 -> 47.0 | Hurt |
| Xavier vs kaiming_uniform prototype initialization (ablation) | PAU · MSR-VTT t2v | R@1: 47.2 -> 48.5 | Improved |
| Xavier vs orthogonal prototype initialization (ablation) | PAU · MSR-VTT v2t | R@1: 47.1 -> 48.3 | Improved |
| train with PAU on small MSVD dataset | CLIP baseline · MSVD v2t | R@1: 60.1 -> 68.9 | Improved |
Closest papers by meaning
- Dynamic Adapter with Semantics Disentangling for Cross-lingual Cross-modal Retrieval (AAAI 2025 · similarity 0.87)
- Fuzzy Multimodal Learning for Trusted Cross-modal Retrieval (CVPR 2025 · similarity 0.87)
- Deep Evidential Hashing for Trustworthy Cross-Modal Retrieval (AAAI 2025 · similarity 0.86)
- WikiDO: A New Benchmark Evaluating Cross-Modal Retrieval for Vision-Language Models (NeurIPS 2024 · similarity 0.86)
- BayesRAG: Probabilistic Mutual Evidence Corroboration for Multimodal Retrieval-Augmented Generation (NeurIPS 2026 · similarity 0.86)
- Beyond Global Similarity: Multi-Conditional Retrieval for Fine-Grained Cross-Modal Understanding (CVPR 2026 · similarity 0.86)
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.
