Abstract
For neuroimaging-based brain disease diagnosis, sparse regression models have proved their effectiveness in handling highdimensional data but with a small number of samples. In this paper, we propose a novel framework that utilizes sparse regression models as target-level representation learner and builds a deep convolutional neural network for clinical decision making. Specifically, we first train multiple sparse regression models, each of which has different values of a regularization control parameter, and use the outputs of the trained regression models as target-level representations. Note that sparse regression models trained with different values of a regularization control parameter potentially select different sets of features from the original ones, thereby they have different powers to predict the response values, i.e., a clinical label and clinical scores in our work. We then construct a deep convolutional neural network by taking the target-level representations as input. Our deep network learns to optimally fuse the predicted response variables, i.e., target-level representations, from the same sparse response model(s) and also those from the neighboring sparse response models. To our best knowledge, this is the first work that systematically integrates sparse regression models with deep neural network. In our experiments with ADNI cohort, we validated the effectiveness of the proposed method by achieving the highest classification accuracies in three different tasks of Alzheimer’s disease and mild cognitive impairment identification.
| Original language | English |
|---|---|
| Title of host publication | Machine Learning in Medical Imaging - 7th International Workshop, MLMI 2016 held in conjunction with MICCAI 2016, Proceedings |
| Publisher | Springer Verlag |
| Pages | 113-121 |
| Number of pages | 9 |
| Volume | 10019 LNCS |
| ISBN (Print) | 9783319471563 |
| DOIs | |
| Publication status | Published - 2016 |
| Event | 7th International Workshop on Machine Learning in Medical Imaging, MLMI 2016 held in conjunction with 19th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2016 - Athens, Greece Duration: 2016 Oct 17 → 2016 Oct 17 |
Publication series
| Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
|---|---|
| Volume | 10019 LNCS |
| ISSN (Print) | 03029743 |
| ISSN (Electronic) | 16113349 |
Other
| Other | 7th International Workshop on Machine Learning in Medical Imaging, MLMI 2016 held in conjunction with 19th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2016 |
|---|---|
| Country | Greece |
| City | Athens |
| Period | 16/10/17 → 16/10/17 |
Fingerprint
ASJC Scopus subject areas
- Theoretical Computer Science
- Computer Science(all)
Cite this
Deep ensemble sparse regression network for alzheimer’s disease diagnosis. / Suk, Heung-Il; Shen, Dinggang.
Machine Learning in Medical Imaging - 7th International Workshop, MLMI 2016 held in conjunction with MICCAI 2016, Proceedings. Vol. 10019 LNCS Springer Verlag, 2016. p. 113-121 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10019 LNCS).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Deep ensemble sparse regression network for alzheimer’s disease diagnosis
AU - Suk, Heung-Il
AU - Shen, Dinggang
PY - 2016
Y1 - 2016
N2 - For neuroimaging-based brain disease diagnosis, sparse regression models have proved their effectiveness in handling highdimensional data but with a small number of samples. In this paper, we propose a novel framework that utilizes sparse regression models as target-level representation learner and builds a deep convolutional neural network for clinical decision making. Specifically, we first train multiple sparse regression models, each of which has different values of a regularization control parameter, and use the outputs of the trained regression models as target-level representations. Note that sparse regression models trained with different values of a regularization control parameter potentially select different sets of features from the original ones, thereby they have different powers to predict the response values, i.e., a clinical label and clinical scores in our work. We then construct a deep convolutional neural network by taking the target-level representations as input. Our deep network learns to optimally fuse the predicted response variables, i.e., target-level representations, from the same sparse response model(s) and also those from the neighboring sparse response models. To our best knowledge, this is the first work that systematically integrates sparse regression models with deep neural network. In our experiments with ADNI cohort, we validated the effectiveness of the proposed method by achieving the highest classification accuracies in three different tasks of Alzheimer’s disease and mild cognitive impairment identification.
AB - For neuroimaging-based brain disease diagnosis, sparse regression models have proved their effectiveness in handling highdimensional data but with a small number of samples. In this paper, we propose a novel framework that utilizes sparse regression models as target-level representation learner and builds a deep convolutional neural network for clinical decision making. Specifically, we first train multiple sparse regression models, each of which has different values of a regularization control parameter, and use the outputs of the trained regression models as target-level representations. Note that sparse regression models trained with different values of a regularization control parameter potentially select different sets of features from the original ones, thereby they have different powers to predict the response values, i.e., a clinical label and clinical scores in our work. We then construct a deep convolutional neural network by taking the target-level representations as input. Our deep network learns to optimally fuse the predicted response variables, i.e., target-level representations, from the same sparse response model(s) and also those from the neighboring sparse response models. To our best knowledge, this is the first work that systematically integrates sparse regression models with deep neural network. In our experiments with ADNI cohort, we validated the effectiveness of the proposed method by achieving the highest classification accuracies in three different tasks of Alzheimer’s disease and mild cognitive impairment identification.
UR - http://www.scopus.com/inward/record.url?scp=84992513254&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84992513254&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-47157-0_14
DO - 10.1007/978-3-319-47157-0_14
M3 - Conference contribution
AN - SCOPUS:84992513254
SN - 9783319471563
VL - 10019 LNCS
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 113
EP - 121
BT - Machine Learning in Medical Imaging - 7th International Workshop, MLMI 2016 held in conjunction with MICCAI 2016, Proceedings
PB - Springer Verlag
ER -