Unverified50% confidenceFactExact time
该算法的损失函数包含三个部分：Focal Loss（正负样本均衡）、相似性损失（相邻点一致性约束）和形状损失（二阶平滑约束）
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026
Sources
深度学习车道线检测：分割转分类实现300+FPS实时推理
bilibili论文发刊罗小黑6/12/2026
Related Claims
Unverified常见的对比学习损失函数包括 NT-Xent Loss 和 Triplet Loss78% similarUnverifiedMRL在训练时同时优化多个维度层级的损失函数，总损失是各粒度损失的加权和69% similarVerified梯度下降通过沿着损失函数梯度的反方向迭代调整权重使预测误差逐步收敛68% similarUnverified对比学习的核心损失函数（如InfoNCE、NT-Xent）本质是拉近同类样本表征距离、推远不同类样本表征距离68% similarUnverifiedGoogLeNet used intermediate classifiers as auxiliary losses to alleviate the vanishing gradient problem.66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/42665API
curl https://kongchang.com/api/v1/knowledge/claims/42665MCP
get_claim(id=42665)