Unverified50% confidenceSolutionExact time
数据增强、标签平滑、适度的权重衰减等技术可以有效压平插值阈值附近的误差尖峰
1
Sources
50%
Confidence
Long-term
Relevance
8/6/2026
First Seen
Sources
Related Claims
Unverified块划分粒度是多目标优化问题：块越小近似精度越高但元数据开销增大且GPU并行效率降低73% similarUnverifiedData normalization makes the loss surface closer to spherical, allowing optimizers to find the optimal solution more efficiently and avoiding oscillation and slow convergence caused by differences in feature scale.72% similarUnverified结构化、上下文丰富的提示不仅提升模型输出质量，也能显著降低安全分类器的误触发率71% similarUnverified应对域偏移的常见策略包括少量标注数据的微调、无监督域适应以及测试时增强70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/693827API
curl https://kongchang.com/api/v1/knowledge/claims/693827MCP
get_claim(id=693827)