arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2402.02012cs.CV
收藏

通过流匹配实现精确知识迁移

Precise Knowledge Transfer via Flow Matching

Mohamed bin Zayed University of Artificial Intelligence(穆罕默德·本·扎耶德人工智能大学);Shanghai AI Laboratory(上海人工智能实验室);Shanghai Jiaotong University(上海交通大学);Tsinghua University(清华大学)

Shitong Shao, Zhiqiang Shen, Linrui Gong, Huanran Chen, Xu Dai

AI总结:

提出基于流匹配的知识迁移框架FM-KT,利用连续归一化流和多步采样实现精确知识迁移,可集成多种蒸馏方法和架构,理论证明其训练目标等价于最小化教师负对数似然上界,并在多个数据集上验证了最先进性能。

摘要:

在本文中,我们提出了一种新颖的知识迁移框架,该框架引入连续归一化流进行渐进式知识转换,并利用多步采样策略实现精确的知识迁移。我们将此框架命名为基于流匹配的知识迁移(FM-KT),它可以与任何形式的基于度量的蒸馏方法(例如 vanilla KD、DKD、PKD 和 DIST)以及任何可用架构的元编码器(例如 CNN、MLP 和 Transformer)集成。通过引入随机插值,FM-KT 可以轻松适应任意噪声调度(例如 VP-ODE、VE-ODE、Rectified flow)以进行归一化流路径估计。我们从理论上证明,FM-KT 的训练目标等价于最小化教师特征图或 logit 负对数似然的上界。此外,FM-KT 可以被视为一种独特的隐式集成方法,从而带来性能提升。通过略微修改 FM-KT 框架,FM-KT 还可以转化为在线蒸馏框架 OFM-KT,并具有可观的性能提升。通过在 CIFAR-100、ImageNet-1k 和 MS-COCO 数据集上的大量实验,我们实证验证了所提出方法在相关比较方法中的可扩展性和最先进性能。

英文摘要:

In this paper, we propose a novel knowledge transfer framework that introduces continuous normalizing flows for progressive knowledge transformation and leverages multi-step sampling strategies to achieve precision knowledge transfer. We name this framework Knowledge Transfer with Flow Matching (FM-KT), which can be integrated with a metric-based distillation method with any form (\textit{e.g.} vanilla KD, DKD, PKD and DIST) and a meta-encoder with any available architecture (\textit{e.g.} CNN, MLP and Transformer). By introducing stochastic interpolants, FM-KD is readily amenable to arbitrary noise schedules (\textit{e.g.}, VP-ODE, VE-ODE, Rectified flow) for normalized flow path estimation. We theoretically demonstrate that the training objective of FM-KT is equivalent to minimizing the upper bound of the teacher feature map or logit negative log-likelihood. Besides, FM-KT can be viewed as a unique implicit ensemble method that leads to performance gains. By slightly modifying the FM-KT framework, FM-KT can also be transformed into an online distillation framework OFM-KT with desirable performance gains. Through extensive experiments on CIFAR-100, ImageNet-1k, and MS-COCO datasets, we empirically validate the scalability and state-of-the-art performance of our proposed methods among relevant comparison approaches.

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。

↑