arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2207.05409cs.CV
收藏

知识浓缩蒸馏

Knowledge Condensation Distillation

Xiamen University(厦门大学);Tencent Youtu Lab(腾讯优图实验室);Hunan University(湖南大学)

Chenxin Li, Mingbao Lin, Zhiyuan Ding, Nie Lin, Yihong Zhuang, Yue Huang, Xinghao Ding, Liujuan Cao

摘要:

知识蒸馏(KD)将高能力教师网络的知识转移到较小的学生身上。现有的方法主要集中在挖掘知识暗示,并将全部知识传递给学生。然而,由于知识在不同的学习阶段对学生表现出不同的价值,从而产生了知识冗余。在本文中,我们提出了知识浓缩蒸馏(KCD)。具体地说,每个样本上的知识值是动态估计的,在此基础上的期望最大化(EM)框架是伪造的迭代浓缩从教师的紧凑的知识集,以指导学生学习。我们的方法很容易建立在现成的KD方法之上,没有额外的训练参数,计算开销可以忽略不计。因此,它提供了一个新的视角KD,其中学生主动识别教师的知识符合自己的能力,可以学会更有效地学习。在标准基准测试上的实验表明,所提出的KCD可以很好地提高学生模型的性能,甚至更高的蒸馏效率。代码可在https://github.com/dzy3/KCD上获得。

英文摘要:

Knowledge Distillation (KD) transfers the knowledge from a high-capacity teacher network to strengthen a smaller student. Existing methods focus on excavating the knowledge hints and transferring the whole knowledge to the student. However, the knowledge redundancy arises since the knowledge shows different values to the student at different learning stages. In this paper, we propose Knowledge Condensation Distillation (KCD). Specifically, the knowledge value on each sample is dynamically estimated, based on which an Expectation-Maximization (EM) framework is forged to iteratively condense a compact knowledge set from the teacher to guide the student learning. Our approach is easy to build on top of the off-the-shelf KD methods, with no extra training parameters and negligible computation overhead. Thus, it presents one new perspective for KD, in which the student that actively identifies teacher's knowledge in line with its aptitude can learn to learn more effectively and efficiently. Experiments on standard benchmarks manifest that the proposed KCD can well boost the performance of student model with even higher distillation efficiency. Code is available at https://github.com/dzy3/KCD.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑