Learning to Learn Kernels with Variational Random Features

Xiantong Zhen, Haoliang Sun, Yingjun Du, Jun Xu, Yilong Yin, Ling Shao, Cees Snoek
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:11409-11419, 2020.

Abstract

We introduce kernels with random Fourier features in the meta-learning framework for few-shot learning. We propose meta variational random features (MetaVRF) to learn adaptive kernels for the base-learner, which is developed in a latent variable model by treating the random feature basis as the latent variable. We formulate the optimization of MetaVRF as a variational inference problem by deriving an evidence lower bound under the meta-learning framework. To incorporate shared knowledge from related tasks, we propose a context inference of the posterior, which is established by an LSTM architecture. The LSTM-based inference network can effectively integrate the context information of previous tasks with task-specific information, generating informative and adaptive features. The learned MetaVRF can produce kernels of high representational power with a relatively low spectral sampling rate and also enables fast adaptation to new tasks. Experimental results on a variety of few-shot regression and classification tasks demonstrate that MetaVRF delivers much better, or at least competitive, performance compared to existing meta-learning alternatives.

Cite this Paper

BibTeX
@InProceedings{pmlr-v119-zhen20a, title = {Learning to Learn Kernels with Variational Random Features}, author = {Zhen, Xiantong and Sun, Haoliang and Du, Yingjun and Xu, Jun and Yin, Yilong and Shao, Ling and Snoek, Cees}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {11409--11419}, year = {2020}, editor = {III, Hal Daumé and Singh, Aarti}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/zhen20a/zhen20a.pdf}, url = {https://proceedings.mlr.press/v119/zhen20a.html}, abstract = {We introduce kernels with random Fourier features in the meta-learning framework for few-shot learning. We propose meta variational random features (MetaVRF) to learn adaptive kernels for the base-learner, which is developed in a latent variable model by treating the random feature basis as the latent variable. We formulate the optimization of MetaVRF as a variational inference problem by deriving an evidence lower bound under the meta-learning framework. To incorporate shared knowledge from related tasks, we propose a context inference of the posterior, which is established by an LSTM architecture. The LSTM-based inference network can effectively integrate the context information of previous tasks with task-specific information, generating informative and adaptive features. The learned MetaVRF can produce kernels of high representational power with a relatively low spectral sampling rate and also enables fast adaptation to new tasks. Experimental results on a variety of few-shot regression and classification tasks demonstrate that MetaVRF delivers much better, or at least competitive, performance compared to existing meta-learning alternatives.} }
Endnote
%0 Conference Paper %T Learning to Learn Kernels with Variational Random Features %A Xiantong Zhen %A Haoliang Sun %A Yingjun Du %A Jun Xu %A Yilong Yin %A Ling Shao %A Cees Snoek %B Proceedings of the 37th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2020 %E Hal Daumé III %E Aarti Singh %F pmlr-v119-zhen20a %I PMLR %P 11409--11419 %U https://proceedings.mlr.press/v119/zhen20a.html %V 119 %X We introduce kernels with random Fourier features in the meta-learning framework for few-shot learning. We propose meta variational random features (MetaVRF) to learn adaptive kernels for the base-learner, which is developed in a latent variable model by treating the random feature basis as the latent variable. We formulate the optimization of MetaVRF as a variational inference problem by deriving an evidence lower bound under the meta-learning framework. To incorporate shared knowledge from related tasks, we propose a context inference of the posterior, which is established by an LSTM architecture. The LSTM-based inference network can effectively integrate the context information of previous tasks with task-specific information, generating informative and adaptive features. The learned MetaVRF can produce kernels of high representational power with a relatively low spectral sampling rate and also enables fast adaptation to new tasks. Experimental results on a variety of few-shot regression and classification tasks demonstrate that MetaVRF delivers much better, or at least competitive, performance compared to existing meta-learning alternatives.
APA
Zhen, X., Sun, H., Du, Y., Xu, J., Yin, Y., Shao, L. & Snoek, C.. (2020). Learning to Learn Kernels with Variational Random Features. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:11409-11419 Available from https://proceedings.mlr.press/v119/zhen20a.html.

Related Material