MemKD：紧凑型循环神经网络记忆保持蒸馏框架

1. 12:00
   arXiv · Machine Learning Theory

   Learning to Remember：为紧凑型循环神经网络蒸馏记忆保持能力

   一篇 arXiv 预印本提出 Memory-Discrepancy Knowledge Distillation（MemKD）框架，用专门损失函数捕捉教师与学生模型在时间序列子序列间的记忆保持差异。实验显示 MemKD 显著优于现有 KD 方法，并可在多种压缩级别下匹配教师模型性能，大幅减少参数量与内存占用且不明显损失精度。

Current heat 9·Comparable peak 10（Oct 7）·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.