Scaling Recurrent Models Via Orthogonal Approximations In Tensor Trains
2019 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2019)(2019)
摘要
Modern deep networks have proven to be very effective for analyzing real world images. However, their application in medical imaging is still in its early stages, primarily due to the large size of three-dimensional images, requiring enormous convolutional or fully connected layers - if we treat an image (and not image patches) as a sample. These issues only compound when the focus moves towards longitudinal analysis of 3D image volumes through recurrent structures, and when a point estimate of model parameters is insufficient in scientific applications where a reliability measure is necessary. Using insights from differential geometry, we adapt the tensor train decomposition to construct networks with significantly fewer parameters, allowing us to train powerful recurrent networks on whole brain image volume sequences. We describe the "orthogonal" tensor train, and demonstrate its ability to express a standard network layer both theoretically and empirically. We show its ability to effectively reconstruct whole brain volumes with faster convergence and stronger confidence intervals compared to the standard tensor train decomposition. We provide code and show experiments on the ADNI dataset using image sequences to regress on a cognition related outcome.
更多查看译文
关键词
recurrent models,orthogonal approximations,tensor trains,modern deep networks,world images,medical imaging,three-dimensional images,enormous convolutional connected layers,fully connected layers,image patches,longitudinal analysis,recurrent structures,point estimate,model parameters,scientific applications,reliability measure,differential geometry,powerful recurrent networks,brain image volume sequences,orthogonal tensor train,standard network layer,brain volumes,standard tensor train decomposition,image sequences
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络