@inproceedings{Baweja:2018, author = {Baweja, C and Glocker, B and Kamnitsas, K}, title = {Towards continual learning in medical imaging}, url = {http://hdl.handle.net/10044/1/66278}, year = {2018} }
TY - CPAPER AB - This work investigates continual learning of two segmentation tasks in brain MRIwith neural networks. To explore in this context the capabilities of current methodsfor countering catastrophic forgetting of the first task when a new one is learned,we investigateelastic weight consolidation[1], a recently proposed method basedon Fisher information, originally evaluated on reinforcement learning of Atarigames. We use it to sequentially learn segmentation of normal brain structures andthen segmentation of white matter lesions. Our findings show this recent methodreduces catastrophic forgetting, while large room for improvement exists in thesechallenging settings for continual learning. AU - Baweja,C AU - Glocker,B AU - Kamnitsas,K PY - 2018/// TI - Towards continual learning in medical imaging UR - http://hdl.handle.net/10044/1/66278 ER -