研究人员开发了一种名为MILD（Margin-based Imbalanced Learning to Defer）的新方法，以解决两阶段延迟学习系统中的专家不平衡问题。该方法将延迟损失优化重新构建为成本敏感学习问题，从而在由于数据不平衡而偏向某些专家的情况下提高了性能。所提出的算法和损失函数在图像分类和大型语言模型（LLM）路由任务中均显示出有效性。 AI
影响 通过解决专家不平衡问题，提高了复杂LLM路由和分类任务的效率和准确性。
排序理由 介绍针对特定机器学习问题的 novel 算法的学术论文。
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
arXiv:2604.27723v1 Announce Type: cross Abstract: Learning algorithms can be significantly improved by routing complex or uncertain inputs to specialized experts, balancing accuracy with computational cost. This approach, known as learning to defer, is essential in domains like n…
Learning algorithms can be significantly improved by routing complex or uncertain inputs to specialized experts, balancing accuracy with computational cost. This approach, known as learning to defer, is essential in domains like natural language generation, medical diagnosis, and…