Incorporating knowledge of plates in batch normalization improves generalization of deep learning for microscopy images

Alexander Lin, Alex Lu
Proceedings of the 17th Machine Learning in Computational Biology meeting, PMLR 200:74-93, 2022.

Abstract

Data collected by high-throughput microscopy experiments are affected by batch effects, stemming from slight technical differences between experimental batches. Batch effects significantly impede machine learning efforts, as models learn spurious technical variation that do not generalize. We introduce batch effects normalization (BEN), a simple method for correcting batch effects that can be applied to any neural network with batch normalization (BN) layers. BEN aligns the concept of a ”batch” in biological experiments with that of a ”batch” in deep learning. During each training step, data points forming the deep learning batch are always sampled from the same experimental batch. This small tweak turns the batch normalization layers into an estimate of the shared batch effects between images, allowing for these technical effects to be standardized out during training and inference. We demonstrate that BEN results in dramatic performance boosts in both supervised and unsupervised learning, leading to state-of-the-art performance on the RxRx1-Wilds benchmark.

Cite this Paper

BibTeX
@InProceedings{pmlr-v200-lin22a, title = {Incorporating knowledge of plates in batch normalization improves generalization of deep learning for microscopy images }, author = {Lin, Alexander and Lu, Alex}, booktitle = {Proceedings of the 17th Machine Learning in Computational Biology meeting}, pages = {74--93}, year = {2022}, editor = {Knowles, David A and Mostafavi, Sara and Lee, Su-In}, volume = {200}, series = {Proceedings of Machine Learning Research}, month = {21--22 Nov}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v200/lin22a/lin22a.pdf}, url = {https://proceedings.mlr.press/v200/lin22a.html}, abstract = {Data collected by high-throughput microscopy experiments are affected by batch effects, stemming from slight technical differences between experimental batches. Batch effects significantly impede machine learning efforts, as models learn spurious technical variation that do not generalize. We introduce batch effects normalization (BEN), a simple method for correcting batch effects that can be applied to any neural network with batch normalization (BN) layers. BEN aligns the concept of a ”batch” in biological experiments with that of a ”batch” in deep learning. During each training step, data points forming the deep learning batch are always sampled from the same experimental batch. This small tweak turns the batch normalization layers into an estimate of the shared batch effects between images, allowing for these technical effects to be standardized out during training and inference. We demonstrate that BEN results in dramatic performance boosts in both supervised and unsupervised learning, leading to state-of-the-art performance on the RxRx1-Wilds benchmark.} }
Endnote
%0 Conference Paper %T Incorporating knowledge of plates in batch normalization improves generalization of deep learning for microscopy images %A Alexander Lin %A Alex Lu %B Proceedings of the 17th Machine Learning in Computational Biology meeting %C Proceedings of Machine Learning Research %D 2022 %E David A Knowles %E Sara Mostafavi %E Su-In Lee %F pmlr-v200-lin22a %I PMLR %P 74--93 %U https://proceedings.mlr.press/v200/lin22a.html %V 200 %X Data collected by high-throughput microscopy experiments are affected by batch effects, stemming from slight technical differences between experimental batches. Batch effects significantly impede machine learning efforts, as models learn spurious technical variation that do not generalize. We introduce batch effects normalization (BEN), a simple method for correcting batch effects that can be applied to any neural network with batch normalization (BN) layers. BEN aligns the concept of a ”batch” in biological experiments with that of a ”batch” in deep learning. During each training step, data points forming the deep learning batch are always sampled from the same experimental batch. This small tweak turns the batch normalization layers into an estimate of the shared batch effects between images, allowing for these technical effects to be standardized out during training and inference. We demonstrate that BEN results in dramatic performance boosts in both supervised and unsupervised learning, leading to state-of-the-art performance on the RxRx1-Wilds benchmark.
APA
Lin, A. & Lu, A.. (2022). Incorporating knowledge of plates in batch normalization improves generalization of deep learning for microscopy images . Proceedings of the 17th Machine Learning in Computational Biology meeting, in Proceedings of Machine Learning Research 200:74-93 Available from https://proceedings.mlr.press/v200/lin22a.html.

Related Material