Batch mode active learning for mitotic phenotypes using conformal prediction

Adam Corrigan, Philip Hopcroft, Ana Narvaez, Claus Bendtsen
Proceedings of the Ninth Symposium on Conformal and Probabilistic Prediction and Applications, PMLR 128:229-243, 2020.

Abstract

Machine learning models are now ubiquitous in all areas of data analysis. As the amount of data generated continues to increase exponentially, the task of annotating sufficient objects with known labels by an expert remains expensive. To mitigate this, active learning approaches attempt to identify those objects whose labels will be most informative. Here, we introduce a batch-based active learning framework in a pooled setting based around conformal predictors. We select objects to add to the labelled observations based on perceived novelty, while mitigating the risks of selecting highly correlated or outlying observations. We compare our approach to classical methods using an example UCI dataset, and demonstrate its application to a pharmaceutically relevant cellular imaging problem for classifying mitotic phenotypes. Our approach facilitates efficient discovery of rare and novel classes within large screening datasets.

Cite this Paper


BibTeX
@InProceedings{pmlr-v128-corrigan20a, title = {Batch mode active learning for mitotic phenotypes using conformal prediction}, author = {Corrigan, Adam and Hopcroft, Philip and Narvaez, Ana and Bendtsen, Claus}, booktitle = {Proceedings of the Ninth Symposium on Conformal and Probabilistic Prediction and Applications}, pages = {229--243}, year = {2020}, editor = {Gammerman, Alexander and Vovk, Vladimir and Luo, Zhiyuan and Smirnov, Evgueni and Cherubin, Giovanni}, volume = {128}, series = {Proceedings of Machine Learning Research}, month = {09--11 Sep}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v128/corrigan20a/corrigan20a.pdf}, url = {https://proceedings.mlr.press/v128/corrigan20a.html}, abstract = {Machine learning models are now ubiquitous in all areas of data analysis. As the amount of data generated continues to increase exponentially, the task of annotating sufficient objects with known labels by an expert remains expensive. To mitigate this, active learning approaches attempt to identify those objects whose labels will be most informative. Here, we introduce a batch-based active learning framework in a pooled setting based around conformal predictors. We select objects to add to the labelled observations based on perceived novelty, while mitigating the risks of selecting highly correlated or outlying observations. We compare our approach to classical methods using an example UCI dataset, and demonstrate its application to a pharmaceutically relevant cellular imaging problem for classifying mitotic phenotypes. Our approach facilitates efficient discovery of rare and novel classes within large screening datasets.} }
Endnote
%0 Conference Paper %T Batch mode active learning for mitotic phenotypes using conformal prediction %A Adam Corrigan %A Philip Hopcroft %A Ana Narvaez %A Claus Bendtsen %B Proceedings of the Ninth Symposium on Conformal and Probabilistic Prediction and Applications %C Proceedings of Machine Learning Research %D 2020 %E Alexander Gammerman %E Vladimir Vovk %E Zhiyuan Luo %E Evgueni Smirnov %E Giovanni Cherubin %F pmlr-v128-corrigan20a %I PMLR %P 229--243 %U https://proceedings.mlr.press/v128/corrigan20a.html %V 128 %X Machine learning models are now ubiquitous in all areas of data analysis. As the amount of data generated continues to increase exponentially, the task of annotating sufficient objects with known labels by an expert remains expensive. To mitigate this, active learning approaches attempt to identify those objects whose labels will be most informative. Here, we introduce a batch-based active learning framework in a pooled setting based around conformal predictors. We select objects to add to the labelled observations based on perceived novelty, while mitigating the risks of selecting highly correlated or outlying observations. We compare our approach to classical methods using an example UCI dataset, and demonstrate its application to a pharmaceutically relevant cellular imaging problem for classifying mitotic phenotypes. Our approach facilitates efficient discovery of rare and novel classes within large screening datasets.
APA
Corrigan, A., Hopcroft, P., Narvaez, A. & Bendtsen, C.. (2020). Batch mode active learning for mitotic phenotypes using conformal prediction. Proceedings of the Ninth Symposium on Conformal and Probabilistic Prediction and Applications, in Proceedings of Machine Learning Research 128:229-243 Available from https://proceedings.mlr.press/v128/corrigan20a.html.

Related Material