ContextNet: Deep learning for Star Galaxy Classification

Noble Kennamer, David Kirkby, Alexander Ihler, Francisco Javier Sanchez-Lopez
Proceedings of the 35th International Conference on Machine Learning, PMLR 80:2582-2590, 2018.

Abstract

We present a framework to compose artificial neural networks in cases where the data cannot be treated as independent events. Our particular motivation is star galaxy classification for ground based optical surveys. Due to a turbulent atmosphere and imperfect instruments, a single image of an astronomical object is not enough to definitively classify it as a star or galaxy. Instead the context of the surrounding objects imaged at the same time need to be considered in order to make an optimal classification. The model we present is divided into three distinct ANNs: one designed to capture local features about each object, the second to compare these features across all objects in an image, and the third to make a final prediction for each object based on the local and compared features. By exploiting the ability to replicate the weights of an ANN, the model can handle an arbitrary and variable number of individual objects embedded in a larger exposure. We train and test our model on simulations of a large up and coming ground based survey, the Large Synoptic Survey Telescope (LSST). We compare to the state of the art approach, showing improved overall performance as well as better performance for a specific class of objects that is important for the LSST.

Cite this Paper


BibTeX
@InProceedings{pmlr-v80-kennamer18a, title = {{C}ontext{N}et: Deep learning for Star Galaxy Classification}, author = {Kennamer, Noble and Kirkby, David and Ihler, Alexander and Sanchez-Lopez, Francisco Javier}, booktitle = {Proceedings of the 35th International Conference on Machine Learning}, pages = {2582--2590}, year = {2018}, editor = {Dy, Jennifer and Krause, Andreas}, volume = {80}, series = {Proceedings of Machine Learning Research}, month = {10--15 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v80/kennamer18a/kennamer18a.pdf}, url = {https://proceedings.mlr.press/v80/kennamer18a.html}, abstract = {We present a framework to compose artificial neural networks in cases where the data cannot be treated as independent events. Our particular motivation is star galaxy classification for ground based optical surveys. Due to a turbulent atmosphere and imperfect instruments, a single image of an astronomical object is not enough to definitively classify it as a star or galaxy. Instead the context of the surrounding objects imaged at the same time need to be considered in order to make an optimal classification. The model we present is divided into three distinct ANNs: one designed to capture local features about each object, the second to compare these features across all objects in an image, and the third to make a final prediction for each object based on the local and compared features. By exploiting the ability to replicate the weights of an ANN, the model can handle an arbitrary and variable number of individual objects embedded in a larger exposure. We train and test our model on simulations of a large up and coming ground based survey, the Large Synoptic Survey Telescope (LSST). We compare to the state of the art approach, showing improved overall performance as well as better performance for a specific class of objects that is important for the LSST.} }
Endnote
%0 Conference Paper %T ContextNet: Deep learning for Star Galaxy Classification %A Noble Kennamer %A David Kirkby %A Alexander Ihler %A Francisco Javier Sanchez-Lopez %B Proceedings of the 35th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2018 %E Jennifer Dy %E Andreas Krause %F pmlr-v80-kennamer18a %I PMLR %P 2582--2590 %U https://proceedings.mlr.press/v80/kennamer18a.html %V 80 %X We present a framework to compose artificial neural networks in cases where the data cannot be treated as independent events. Our particular motivation is star galaxy classification for ground based optical surveys. Due to a turbulent atmosphere and imperfect instruments, a single image of an astronomical object is not enough to definitively classify it as a star or galaxy. Instead the context of the surrounding objects imaged at the same time need to be considered in order to make an optimal classification. The model we present is divided into three distinct ANNs: one designed to capture local features about each object, the second to compare these features across all objects in an image, and the third to make a final prediction for each object based on the local and compared features. By exploiting the ability to replicate the weights of an ANN, the model can handle an arbitrary and variable number of individual objects embedded in a larger exposure. We train and test our model on simulations of a large up and coming ground based survey, the Large Synoptic Survey Telescope (LSST). We compare to the state of the art approach, showing improved overall performance as well as better performance for a specific class of objects that is important for the LSST.
APA
Kennamer, N., Kirkby, D., Ihler, A. & Sanchez-Lopez, F.J.. (2018). ContextNet: Deep learning for Star Galaxy Classification. Proceedings of the 35th International Conference on Machine Learning, in Proceedings of Machine Learning Research 80:2582-2590 Available from https://proceedings.mlr.press/v80/kennamer18a.html.

Related Material