Research map / Label-efficient and robust learning / Domain adaptation and generalization
Few-shot Image Generation via Adaptation-Aware Kernel Modulation
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 importance probing (select kernels to modulate) vs modulating all kernels (ablation) | AdAM · FFHQ->AFHQ-Cat | FID: 82.41 -> 58.07 | Helped |
| add importance probing vs modulating all kernels (ablation) | AdAM · FFHQ->Babies | FID: 54.46 -> 48.83 | Helped |
| modulate important kernels with KML instead of freezing them (ablation) | AdAM · FFHQ->AFHQ-Cat | FID: 61.60 -> 58.07 | Helped |
| modulate important kernels with KML instead of freezing them (ablation) | AdAM · FFHQ->Babies | FID: 50.81 -> 48.83 | Helped |
| add importance probing to EWC (ablation) | EWC · FFHQ->Babies | FID: 79.93 -> 70.80 | Helped |
| add importance probing to EWC (ablation) | EWC · FFHQ->AFHQ-Cat | FID: 74.61 -> 66.35 | Helped |
| add importance probing to AdaFM (ablation) | AdaFM · FFHQ->Babies | FID: 62.90 -> 55.64 | Helped |
| add importance probing to AdaFM (ablation) | AdaFM · FFHQ->AFHQ-Cat | FID: 64.44 -> 60.04 | Helped |
| use source-only knowledge-preserving methods (EWC, CDC, DCL) instead of plain fine-tuning | TGAN baseline · FFHQ->AFHQ-Cat | FID: 64.68 -> 176.21 (CDC) | Hurt |
| use CDC/DCL instead of plain fine-tuning on close-proximity target | TGAN baseline · FFHQ->Babies | FID: 101.58 -> 56.48 (DCL) | Helped |
| replace DCL with AdAM on distant target | DCL · FFHQ->AFHQ-Wild | FID: 115.93 -> 36.87 | Helped |
| add importance probing to AdAM (ablation) | AdAM · FFHQ->Babies | Intra-LPIPS: 0.613 -> 0.590 | Hurt |
Closest papers by meaning
- Few-Shot Image Generation via Cross-Domain Correspondence (CVPR 2021 · similarity 0.91)
- A Closer Look at Few-Shot Image Generation (CVPR 2022 · similarity 0.91)
- Uni-DAD: Unified Distillation and Adaptation of Diffusion Models for Few-step Few-shot Image Generation (CVPR 2026 · similarity 0.90)
- Exploring Incompatible Knowledge Transfer in Few-Shot Image Generation (CVPR 2023 · similarity 0.90)
- WeditGAN: Few-Shot Image Generation via Latent Space Relocation (AAAI 2024 · similarity 0.87)
- ProKeR: A Kernel Perspective on Few-Shot Adaptation of Large Vision-Language Models (CVPR 2025 · 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.