Research map / Vision-language and multimodal / Cross-modal learning
Hierarchical Encoding Tree with Modality Mixup for Cross-modal Hashing
Zhiping Xiao, Junyu Luo, Hang Zhou, Yusheng Zhao, Xiao Luo, Pengyun Wang, Wei Ju, Siyu Heng, Ming Zhang
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 intra-modal KNN samples to text-image pairs (ablation) | HINT V1 · MIRFlickr-25K I2T 128-bit | MAP (%): 73.2 -> 73.8 | Improved |
| add hierarchical encoding tree sample selection (ablation) | HINT V2 · MIRFlickr-25K I2T 128-bit | MAP (%): 73.8 -> 74.2 | Improved |
| add hierarchical encoding tree sample selection (ablation) | HINT V2 · MS-COCO I2T 128-bit | MAP (%): 59.1 -> 60.0 | Improved |
| add curriculum-based progressive mixup (ablation) | HINT V3 · NUS-WIDE I2T 128-bit | MAP (%): 65.9 -> 67.0 | Improved |
| add proxy-based consistency learning (ablation) | HINT V4 · MIRFlickr-25K T2I 128-bit | MAP (%): 74.1 -> 74.6 | Improved |
| add proxy-based consistency learning (ablation) | HINT V4 · MS-COCO I2T 128-bit | MAP (%): 60.7 -> 61.1 | Improved |
| use cosine instead of L1 similarity for tree construction (ablation) | HINT · MIRFlickr-25K I2T | MAP (%): 73.5 -> 75.5 | Improved |
| use cosine instead of L2 similarity for tree construction (ablation) | HINT · MIRFlickr-25K I2T | MAP (%): 74.8 -> 75.5 | Improved |
| iteratively update the tree every 5 epochs (ablation) | HINT · MIRFlickr-25K I2T / MS-COCO I2T | MAP (%): 75.5 -> 75.0; 61.1 -> 60.5 | Hurt |
| iteratively update the tree every 5 epochs (ablation) | HINT · MIRFlickr-25K T2I | MAP (%): 74.6 -> 74.7 | No clear effect |
| increase K from 3 to 5 neighbours (ablation) | HINT · MIRFlickr-25K | MAP | Hurt |
| corrupt 10% of pairs: HINT vs DEMO | DEMO · MIRFlickr-25K I2T 128-bit noisy | MAP (%): 71.4 -> 74.4 | Improved |
Closest papers by meaning
- UGGRH: Unsupervised Generative Completion and Graph-attention Refinement for Incomplete Cross-modal Hashing (NeurIPS 2026 · similarity 0.91)
- Dynamic Masking and Auxiliary Hash Learning for Enhanced Cross-Modal Retrieval (NeurIPS 2025 · similarity 0.91)
- Decoupled Prototype Contrastive Alignment Hashing for Cross-Modal Retrieval (NeurIPS 2026 · similarity 0.90)
- Lightweight Contrastive Distilled Hashing for Online Cross-modal Retrieval (AAAI 2025 · similarity 0.90)
- Vision-guided Text Mining for Unsupervised Cross-modal Hashing with Community Similarity Quantization (AAAI 2025 · similarity 0.89)
- An End-To-End Graph Attention Network Hashing for Cross-Modal Retrieval (NeurIPS 2024 · similarity 0.89)
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.
