Unverified85% confidenceFactTime unknown
小样本学习（Few-Shot Learning）的主流方法包括基于度量学习的Siamese Network、基于元学习的MAML，以及结合预训练大模型的提示微调
1
Sources
85%
Confidence
Long-term
Relevance
5/31/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedFew-Shot Learning refers to providing a small number of examples (typically 2–8) in the prompt, allowing the model to complete new tasks through analogical reasoning.76% similarUnverified领域自适应作为迁移学习子领域，常用方法包括对抗训练（DANN）、最大均值差异最小化（MMD）以及基于元学习框架（如MAML）的少样本适应75% similarUnverifiedWarmup策略在训练初期使用极小学习率，待参数进入合理范围后再提升，被大型语言模型训练广泛采用74% similarUnverifiedWhen a Skill is presented in the form of a few-shot example set, it essentially leverages the model's In-Context Learning capability without modifying the model weights.74% similarVerifiedProviding 3–5 high-quality examples (Few-shot Learning) can generally significantly improve the model's instruction-following accuracy.74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/15839API
curl https://kongchang.com/api/v1/knowledge/claims/15839MCP
get_claim(id=15839)