PACOH: Bayes-Optimal Meta-Learning with PAC-Guarantees

Jonas Rothfuss, Vincent Fortuin, Martin Josifoski, Andreas Krause
Proceedings of the 38th International Conference on Machine Learning, PMLR 139:9116-9126, 2021.

Abstract

Meta-learning can successfully acquire useful inductive biases from data. Yet, its generalization properties to unseen learning tasks are poorly understood. Particularly if the number of meta-training tasks is small, this raises concerns about overfitting. We provide a theoretical analysis using the PAC-Bayesian framework and derive novel generalization bounds for meta-learning. Using these bounds, we develop a class of PAC-optimal meta-learning algorithms with performance guarantees and a principled meta-level regularization. Unlike previous PAC-Bayesian meta-learners, our method results in a standard stochastic optimization problem which can be solved efficiently and scales well.When instantiating our PAC-optimal hyper-posterior (PACOH) with Gaussian processes and Bayesian Neural Networks as base learners, the resulting methods yield state-of-the-art performance, both in terms of predictive accuracy and the quality of uncertainty estimates. Thanks to their principled treatment of uncertainty, our meta-learners can also be successfully employed for sequential decision problems.

Cite this Paper


BibTeX
@InProceedings{pmlr-v139-rothfuss21a, title = {PACOH: Bayes-Optimal Meta-Learning with PAC-Guarantees}, author = {Rothfuss, Jonas and Fortuin, Vincent and Josifoski, Martin and Krause, Andreas}, booktitle = {Proceedings of the 38th International Conference on Machine Learning}, pages = {9116--9126}, year = {2021}, editor = {Meila, Marina and Zhang, Tong}, volume = {139}, series = {Proceedings of Machine Learning Research}, month = {18--24 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v139/rothfuss21a/rothfuss21a.pdf}, url = {https://proceedings.mlr.press/v139/rothfuss21a.html}, abstract = {Meta-learning can successfully acquire useful inductive biases from data. Yet, its generalization properties to unseen learning tasks are poorly understood. Particularly if the number of meta-training tasks is small, this raises concerns about overfitting. We provide a theoretical analysis using the PAC-Bayesian framework and derive novel generalization bounds for meta-learning. Using these bounds, we develop a class of PAC-optimal meta-learning algorithms with performance guarantees and a principled meta-level regularization. Unlike previous PAC-Bayesian meta-learners, our method results in a standard stochastic optimization problem which can be solved efficiently and scales well.When instantiating our PAC-optimal hyper-posterior (PACOH) with Gaussian processes and Bayesian Neural Networks as base learners, the resulting methods yield state-of-the-art performance, both in terms of predictive accuracy and the quality of uncertainty estimates. Thanks to their principled treatment of uncertainty, our meta-learners can also be successfully employed for sequential decision problems.} }
Endnote
%0 Conference Paper %T PACOH: Bayes-Optimal Meta-Learning with PAC-Guarantees %A Jonas Rothfuss %A Vincent Fortuin %A Martin Josifoski %A Andreas Krause %B Proceedings of the 38th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2021 %E Marina Meila %E Tong Zhang %F pmlr-v139-rothfuss21a %I PMLR %P 9116--9126 %U https://proceedings.mlr.press/v139/rothfuss21a.html %V 139 %X Meta-learning can successfully acquire useful inductive biases from data. Yet, its generalization properties to unseen learning tasks are poorly understood. Particularly if the number of meta-training tasks is small, this raises concerns about overfitting. We provide a theoretical analysis using the PAC-Bayesian framework and derive novel generalization bounds for meta-learning. Using these bounds, we develop a class of PAC-optimal meta-learning algorithms with performance guarantees and a principled meta-level regularization. Unlike previous PAC-Bayesian meta-learners, our method results in a standard stochastic optimization problem which can be solved efficiently and scales well.When instantiating our PAC-optimal hyper-posterior (PACOH) with Gaussian processes and Bayesian Neural Networks as base learners, the resulting methods yield state-of-the-art performance, both in terms of predictive accuracy and the quality of uncertainty estimates. Thanks to their principled treatment of uncertainty, our meta-learners can also be successfully employed for sequential decision problems.
APA
Rothfuss, J., Fortuin, V., Josifoski, M. & Krause, A.. (2021). PACOH: Bayes-Optimal Meta-Learning with PAC-Guarantees. Proceedings of the 38th International Conference on Machine Learning, in Proceedings of Machine Learning Research 139:9116-9126 Available from https://proceedings.mlr.press/v139/rothfuss21a.html.

Related Material