Deep Multi-Fidelity Active Learning of High-dimensional Outputs
Deep Multi-Fidelity Active Learning of High-dimensional Outputs
复制标题
高维输出的深度多保真主动学习
DOI:
--
复制
发表时间:
2020
期刊:
影响因子:
--
通讯作者:
Shandian Zhe
中科院分区:
文献类型:
--
作者:
Shibo Li;R. Kirby;Shandian Zhe
Many applications, such as in physical simulation and engineering design, demand we estimate functions with high-dimensional outputs. The training examples can be collected with different fidelities to allow a cost/accuracy trade-off. In this paper, we consider the active learning task that identifies both the fidelity and input to query new training examples so as to achieve the best benefit-cost ratio. To this end, we propose DMFAL, a Deep Multi-Fidelity Active Learning approach. We first develop a deep neural network-based multi-fidelity model for learning with high-dimensional outputs, which can flexibly, efficiently capture all kinds of complex relationships across the outputs and fidelities to improve prediction. We then propose a mutual information-based acquisition function that extends the predictive entropy principle. To overcome the computational challenges caused by large output dimensions, we use multi-variate Delta's method and moment-matching to estimate the output posterior, and Weinstein-Aronszajn identity to calculate and optimize the acquisition function. The computation is tractable, reliable and efficient. We show the advantage of our method in several applications of computational physics and engineering design.
DOI:
--
发表时间:
2018-11
期刊:
ArXiv
影响因子:
--
作者:
Jialin Song;Yuxin Chen;Yisong Yue
通讯作者:
Jialin Song;Yuxin Chen;Yisong Yue
DOI:
--
发表时间:
2020-06
期刊:
ArXiv
影响因子:
--
作者:
Zheng Wang;Wei W. Xing;R. Kirby;Shandian Zhe
通讯作者:
Zheng Wang;Wei W. Xing;R. Kirby;Shandian Zhe
DOI:
--
发表时间:
2019-06
期刊:
ArXiv
影响因子:
--
作者:
Andreas Kirsch;Joost R. van Amersfoort;Y. Gal
通讯作者:
Andreas Kirsch;Joost R. van Amersfoort;Y. Gal