研究库 论文知识库

内容库 / 论文

arXiv:2609.22323 · 评测基准

ALPINE: Adaptive Localization for Parameter- and Sample-Efficient Few-Shot Learning

ALPINE: Adaptive Localization for Parameter- and Sample-Efficient Few-Shot Learning

类型：arxiv
标识：2609.22323
链接：https://arxiv.org/abs/2609.22323
主分类：evaluation
形态：method
TLDR：Few-shot learning research is predominantly evaluated on accuracy alone, with limited attention to the parameter and training-sample budgets required to reach that accuracy - a real constraint for practitioners without large-scale compute. We present an ultra-lightweight (22,249-34,917 parameter) spatial-relational architecture for few-shot image classification that combines fixed Gabor edge-energy guidance with a windowed, content-adaptive patch locator. Under a strictly matched, iso-episode-budget protocol (250 meta-training episodes, 5 canonical seeds, 600 evaluation episodes per seed), our
副分类：engineering
待LLM分类：否
来源文件：
/inbox/tom/_candidates/2026-09-23-agent-rag-longcontext-candidates.json

相关 · 按语义匹配

笔记精读与批判 · BDH-CQ · In-Context Learning with Recurrent Latent Reasoning

仓库 sisinflab/adversarial-recommender-systems-survey

笔记 Tom 文献雷达 · Agent RAG LongContext · 2026-08-12 20:40

仓库 Kobaayyy/Awesome-CVPR2026-CVPR2025-ICCV2025-CVPR2024-ECCV2026-ECCV2024-AIGC

笔记 2026-07-03 轻量精读：Qwen-Image-Agent × The Verification Horizon

仓库 ChlorineHi/ST-Path-Survey

arXiv

2609.22323

TLDR

Few-shot learning research is predominantly evaluated on accuracy alone, with limited attention to the parameter and training-sample budgets required to reach that accuracy - a real constraint for practitioners without large-scale compute. We present an ultra-lightweight (22,249-34,917 parameter) spatial-relational architecture for few-shot image classification that combines fixed Gabor edge-energy guidance with a windowed, content-adaptive patch locator. Under a strictly matched, iso-episode-budget protocol (250 meta-training episodes, 5 canonical seeds, 600 evaluation episodes per seed), our

arXiv ↗ 中文全文