Research map / Label-efficient and robust learning / Self-supervised and contrastive
Geometry-Aware Contrastive Learning for Few-Shot Automatic Modulation Recognition
guanqun zhao, Liu Yitong, Jiaxuan Fang, Yufei Mao, Hongwen Yang
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 Dynamic-Consistency (VAA + semantic consistency loss) (ablation) | DyCo-CL · RML2016.10a N=1 | accuracy (%): 34.94 -> 43.84 | Improved |
| use Signal-Adaptive Swin backbone instead of ResNet18 (ablation) | DyCo-CL · RML2016.10a N=1 | accuracy (%): 39.45 -> 43.84 | Improved |
| use adaptive Stage I prior encoding instead of concatenation (ablation) | DyCo-CL · RML2016.10a N=1 | accuracy (%): 39.47 -> 43.84 | Improved |
| use physics-aware transformer fusion Stage II instead of concatenation (ablation) | DyCo-CL · RML2016.10a N=1 | accuracy (%): 40.12 -> 43.84 | Improved |
| use hierarchical hybrid fusion instead of concatenation at all stages (ablation) | DyCo-CL · RML2016.10a N=1 | accuracy (%): 38.32 -> 43.84 | Improved |
| use DyCo-CL instead of best prior SOTA | DyCo-CL vs CMSSAN · RML2016.10a N=1 | accuracy (%): 36.92% -> 43.84% | Improved |
| use DyCo-CL instead of SSCL-AMC | DyCo-CL vs SSCL-AMC · RML2016.10a N=2 | accuracy (%): 41.05% -> 46.58% | Improved |
| use DyCo-CL on larger dataset | DyCo-CL vs SSCL-AMC · RML2018.01a 10-shot | accuracy (%): 32.87% -> 37.96% | Improved |
| use DyCo-CL instead of SOTA at SNR 10 dB | DyCo-CL · RML2016.10a N=1, 10 dB | accuracy: +7.7% | Improved |
| use DyCo-CL instead of SSCL-AMC for efficiency | DyCo-CL vs SSCL-AMC · compute | FLOPs: over 2.5x fewer FLOPs, 3x throughput | Improved |
Closest papers by meaning
- SDGMNet: Statistic-Based Dynamic Gradient Modulation for Local Descriptor Learning (AAAI 2024 · similarity 0.82)
- Modality-Aware SAM: Sharpness-Aware-Minimization Driven Gradient Modulation for Harmonized Multimodal Learning (NeurIPS 2025 · similarity 0.82)
- Contrastive Consolidation of Top-Down Modulations Achieves Sparsely Supervised Continual Learning (NeurIPS 2025 · similarity 0.81)
- Structuring Representation Geometry with Rotationally Equivariant Contrastive Learning (ICLR 2024 · similarity 0.81)
- MM-Mixing: Multi-Modal Mixing Alignment for 3D Understanding (AAAI 2025 · similarity 0.80)
- Contrastive-Equivariant Self-Supervised Learning Improves Alignment with Primate Visual Area IT (NeurIPS 2024 · similarity 0.80)
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.
