Uncertainty-Aware Learning from Multi-Expert Interval Target

1. 12:00
   arXiv · Machine Learning Theory

   Uncertainty-Aware Learning from Multi-Expert Interval Targets

   提出一种不确定性感知的多专家区间目标学习方法，保留个体专家区间并用 Beta 混合分布建模，通过 Cramér 距离目标训练。将预测不确定性分解为 within-component、between-component 和模型不确定性，并引入分解匹配以对齐预测分量与标签侧来源。在海冰浓度数据上 MAE 较硬标签降低 31%，优于聚合、区间分布和区间回归基线。

Current heat 9·Comparable peak 10（Oct 2）·Comparable change over 24 hours –

The trend compares only the same participants observed continuously; its range may be smaller than the current heat count. Move or click on the chart to inspect hourly heat; use the left and right arrow keys to switch.