Research map / Vision-language and multimodal / Cross-modal learning
Adaptive Cross-Modal Few-shot Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 adaptive semantic-visual prototype mixing | TADAM · miniImageNet, 5-way 1-shot | accuracy: 58.56 ± 0.39% -> 65.30 ± 0.49% | Helped |
| add adaptive semantic-visual prototype mixing | TADAM · tieredImageNet, 5-way 1-shot | accuracy: 62.13 ± 0.31% -> 69.08 ± 0.47% | Helped |
| add adaptive semantic-visual prototype mixing | ProtoNets++ · miniImageNet, 5-way 1-shot | accuracy: 56.52 ± 0.45% -> 65.21 ± 0.30% | Helped |
| add adaptive semantic-visual prototype mixing | ProtoNets++ · tieredImageNet, 5-way 1-shot | accuracy: 58.47 ± 0.64% -> 67.23 ± 0.34% | Helped |
| condition the mixing network on transformed semantic features rather than raw label embeddings (ablation) | AM3 with ProtoNets++ · miniImageNet, 1-shot | accuracy: 61.23 -> 65.21 | Helped |
| condition the mixing network on transformed semantic features rather than raw label embeddings (ablation) | AM3 with ProtoNets++ · miniImageNet, 5-shot | accuracy: 74.77 -> 75.20 | Helped |
| condition the mixing network on transformed semantic features rather than raw label embeddings (ablation) | AM3-TADAM · miniImageNet, 1-shot | accuracy: 57.47 -> 65.30 | Helped |
| condition the mixing network on transformed semantic features rather than raw label embeddings (ablation) | AM3-TADAM · miniImageNet, 5-shot | accuracy: 72.27 -> 78.10 | Helped |
| condition the mixing network on transformed semantic features rather than visual prototypes (ablation) | AM3 with ProtoNets++ · miniImageNet, 1-shot | accuracy: 64.48 -> 65.21 | Helped |
| condition the mixing network on transformed semantic features rather than visual prototypes (ablation) | AM3 with ProtoNets++ · miniImageNet, 5-shot | accuracy: 74.80 -> 75.20 | Helped |
| condition the mixing network on transformed semantic features rather than visual prototypes (ablation) | AM3-TADAM · miniImageNet, 1-shot | accuracy: 64.93 -> 65.30 | Helped |
| condition the mixing network on transformed semantic features rather than visual prototypes (ablation) | AM3-TADAM · miniImageNet, 5-shot | accuracy: 77.60 -> 78.10 | Helped |
Closest papers by meaning
- Simple Semantic-Aided Few-Shot Learning (CVPR 2024 · similarity 0.90)
- Pushing the Limit of Fine-Tuning for Few-Shot Learning: Where Feature Reusing Meets Cross-Scale Attention (AAAI 2024 · similarity 0.90)
- Text Augmented Correlation Transformer For Few-shot Classification & Segmentation (CVPR 2025 · similarity 0.89)
- VT-FSL: Bridging Vision and Text with LLMs for Few-Shot Learning (NeurIPS 2025 · similarity 0.89)
- Making Large Vision Language Models to Be Good Few-Shot Learners (AAAI 2025 · similarity 0.88)
- Mind the Gap Between Prototypes and Images in Cross-domain Finetuning (NeurIPS 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.