Adjusting Model Size in Continual Gaussian Processes: How Big is Big Enough?

Guiomar Pescador-Barrios, Sarah Lucie Filippi, Mark Van Der Wilk
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:48974-49000, 2025.

Abstract

Many machine learning models require setting a parameter that controls their size before training, e.g. number of neurons in DNNs, or inducing points in GPs. Increasing capacity typically improves performance until all the information from the dataset is captured. After this point, computational cost keeps increasing, without improved performance. This leads to the question "How big is big enough?" We investigate this problem for Gaussian processes (single-layer neural networks) in continual learning. Here, data becomes available incrementally, and the final dataset size will therefore not be known before training, preventing the use of heuristics for setting a fixed model size. We develop a method to automatically adjust model size while maintaining near-optimal performance. Our experimental procedure follows the constraint that any hyperparameters must be set without seeing dataset properties, and we show that our method performs well across diverse datasets without the need to adjust its hyperparameter, showing it requires less tuning than others.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-pescador-barrios25a, title = {Adjusting Model Size in Continual {G}aussian Processes: How Big is Big Enough?}, author = {Pescador-Barrios, Guiomar and Filippi, Sarah Lucie and Van Der Wilk, Mark}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {48974--49000}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/pescador-barrios25a/pescador-barrios25a.pdf}, url = {https://proceedings.mlr.press/v267/pescador-barrios25a.html}, abstract = {Many machine learning models require setting a parameter that controls their size before training, e.g. number of neurons in DNNs, or inducing points in GPs. Increasing capacity typically improves performance until all the information from the dataset is captured. After this point, computational cost keeps increasing, without improved performance. This leads to the question "How big is big enough?" We investigate this problem for Gaussian processes (single-layer neural networks) in continual learning. Here, data becomes available incrementally, and the final dataset size will therefore not be known before training, preventing the use of heuristics for setting a fixed model size. We develop a method to automatically adjust model size while maintaining near-optimal performance. Our experimental procedure follows the constraint that any hyperparameters must be set without seeing dataset properties, and we show that our method performs well across diverse datasets without the need to adjust its hyperparameter, showing it requires less tuning than others.} }
Endnote
%0 Conference Paper %T Adjusting Model Size in Continual Gaussian Processes: How Big is Big Enough? %A Guiomar Pescador-Barrios %A Sarah Lucie Filippi %A Mark Van Der Wilk %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-pescador-barrios25a %I PMLR %P 48974--49000 %U https://proceedings.mlr.press/v267/pescador-barrios25a.html %V 267 %X Many machine learning models require setting a parameter that controls their size before training, e.g. number of neurons in DNNs, or inducing points in GPs. Increasing capacity typically improves performance until all the information from the dataset is captured. After this point, computational cost keeps increasing, without improved performance. This leads to the question "How big is big enough?" We investigate this problem for Gaussian processes (single-layer neural networks) in continual learning. Here, data becomes available incrementally, and the final dataset size will therefore not be known before training, preventing the use of heuristics for setting a fixed model size. We develop a method to automatically adjust model size while maintaining near-optimal performance. Our experimental procedure follows the constraint that any hyperparameters must be set without seeing dataset properties, and we show that our method performs well across diverse datasets without the need to adjust its hyperparameter, showing it requires less tuning than others.
APA
Pescador-Barrios, G., Filippi, S.L. & Van Der Wilk, M.. (2025). Adjusting Model Size in Continual Gaussian Processes: How Big is Big Enough?. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:48974-49000 Available from https://proceedings.mlr.press/v267/pescador-barrios25a.html.

Related Material