Multi-task Learning with Weak Class Labels: Leveraging iEEG to Detect Cortical Lesions in Cryptogenic Epilepsy

Bilal Ahmed, Thomas Thesen, Karen Blackmon, Ruben Kuzniecky, Orrin Devinsky, Jennifer Dy, Carla Brodley
; Proceedings of the 1st Machine Learning for Healthcare Conference, PMLR 56:115-133, 2016.

Abstract

Multi-task learning (MTL) is useful for domains in which data originates from multiple sources that are individually under-sampled. MTL methods are able to learn classification models that have higher performance as compared to learning a single model by aggregating all the data together or learning a separate model for each data source. The performance of these methods relies on label accuracy. We develop two models that address the problem of multitask learning when the training data has imprecise labels. We apply these methods to the task of detecting abnormal cortical regions in the MRIs of patients suffering from epilepsy whose MRI were deemed normal by neuroradiologists. We use the results of intracranial-EEG exam as an auxiliary source of supervision. The proposed methods successfully detect abnormal regions for all patients in our sample and achieve higher performance as compared to other methods.

Cite this Paper


BibTeX
@InProceedings{pmlr-v56-Ahmed16, title = {Multi-task Learning with Weak Class Labels: Leveraging iEEG to Detect Cortical Lesions in Cryptogenic Epilepsy}, author = {Bilal Ahmed and Thomas Thesen and Karen Blackmon and Ruben Kuzniecky and Orrin Devinsky and Jennifer Dy and Carla Brodley}, booktitle = {Proceedings of the 1st Machine Learning for Healthcare Conference}, pages = {115--133}, year = {2016}, editor = {Finale Doshi-Velez and Jim Fackler and David Kale and Byron Wallace and Jenna Wiens}, volume = {56}, series = {Proceedings of Machine Learning Research}, address = {Northeastern University, Boston, MA, USA}, month = {18--19 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v56/Ahmed16.pdf}, url = {http://proceedings.mlr.press/v56/Ahmed16.html}, abstract = {Multi-task learning (MTL) is useful for domains in which data originates from multiple sources that are individually under-sampled. MTL methods are able to learn classification models that have higher performance as compared to learning a single model by aggregating all the data together or learning a separate model for each data source. The performance of these methods relies on label accuracy. We develop two models that address the problem of multitask learning when the training data has imprecise labels. We apply these methods to the task of detecting abnormal cortical regions in the MRIs of patients suffering from epilepsy whose MRI were deemed normal by neuroradiologists. We use the results of intracranial-EEG exam as an auxiliary source of supervision. The proposed methods successfully detect abnormal regions for all patients in our sample and achieve higher performance as compared to other methods.} }
Endnote
%0 Conference Paper %T Multi-task Learning with Weak Class Labels: Leveraging iEEG to Detect Cortical Lesions in Cryptogenic Epilepsy %A Bilal Ahmed %A Thomas Thesen %A Karen Blackmon %A Ruben Kuzniecky %A Orrin Devinsky %A Jennifer Dy %A Carla Brodley %B Proceedings of the 1st Machine Learning for Healthcare Conference %C Proceedings of Machine Learning Research %D 2016 %E Finale Doshi-Velez %E Jim Fackler %E David Kale %E Byron Wallace %E Jenna Wiens %F pmlr-v56-Ahmed16 %I PMLR %J Proceedings of Machine Learning Research %P 115--133 %U http://proceedings.mlr.press %V 56 %W PMLR %X Multi-task learning (MTL) is useful for domains in which data originates from multiple sources that are individually under-sampled. MTL methods are able to learn classification models that have higher performance as compared to learning a single model by aggregating all the data together or learning a separate model for each data source. The performance of these methods relies on label accuracy. We develop two models that address the problem of multitask learning when the training data has imprecise labels. We apply these methods to the task of detecting abnormal cortical regions in the MRIs of patients suffering from epilepsy whose MRI were deemed normal by neuroradiologists. We use the results of intracranial-EEG exam as an auxiliary source of supervision. The proposed methods successfully detect abnormal regions for all patients in our sample and achieve higher performance as compared to other methods.
RIS
TY - CPAPER TI - Multi-task Learning with Weak Class Labels: Leveraging iEEG to Detect Cortical Lesions in Cryptogenic Epilepsy AU - Bilal Ahmed AU - Thomas Thesen AU - Karen Blackmon AU - Ruben Kuzniecky AU - Orrin Devinsky AU - Jennifer Dy AU - Carla Brodley BT - Proceedings of the 1st Machine Learning for Healthcare Conference PY - 2016/12/10 DA - 2016/12/10 ED - Finale Doshi-Velez ED - Jim Fackler ED - David Kale ED - Byron Wallace ED - Jenna Wiens ID - pmlr-v56-Ahmed16 PB - PMLR SP - 115 DP - PMLR EP - 133 L1 - http://proceedings.mlr.press/v56/Ahmed16.pdf UR - http://proceedings.mlr.press/v56/Ahmed16.html AB - Multi-task learning (MTL) is useful for domains in which data originates from multiple sources that are individually under-sampled. MTL methods are able to learn classification models that have higher performance as compared to learning a single model by aggregating all the data together or learning a separate model for each data source. The performance of these methods relies on label accuracy. We develop two models that address the problem of multitask learning when the training data has imprecise labels. We apply these methods to the task of detecting abnormal cortical regions in the MRIs of patients suffering from epilepsy whose MRI were deemed normal by neuroradiologists. We use the results of intracranial-EEG exam as an auxiliary source of supervision. The proposed methods successfully detect abnormal regions for all patients in our sample and achieve higher performance as compared to other methods. ER -
APA
Ahmed, B., Thesen, T., Blackmon, K., Kuzniecky, R., Devinsky, O., Dy, J. & Brodley, C.. (2016). Multi-task Learning with Weak Class Labels: Leveraging iEEG to Detect Cortical Lesions in Cryptogenic Epilepsy. Proceedings of the 1st Machine Learning for Healthcare Conference, in PMLR 56:115-133

Related Material