#### DMCA

## To cite this version: (2014)

### Citations

1049 | A fast iterative shrinkage-thresholding algorithm for linear inverse problems
- Beck, Teboulle
(Show Context)
Citation Context ...z gradient, whilst the `1 –though nonsmooth– is proximable1 by means of the soft-thesholding operator [11]. Thus problem (2) is amenable to the FISTA (Fast Iterative Shrinkage-Thresholding Algorithm) =-=[12]-=-, with a provable O(1/√) convergence rate. Our implementation of FISTA uses technical recommendations (line-searching, parametrization, etc.) which were provided in [13], in the context of TV-L1 [2],... |

742 | An iterative thresholding algorithm for linear inverse problems with a sparsity constraint - Daubechies, Defrise, et al. - 2004 |

436 | Distributed and overlapping representations of faces and objects in ventral temporal cortex
- Haxby, Gobbini, et al.
- 2001
(Show Context)
Citation Context ...regions obtained here for the 10th and 20th screening-percentiles match pretty well the results obtained in [3] with their TV-L1 penalty. (b): Face vs House contrast of the visual recognition dataset =-=[9]-=-. Weights maps obtained for the GraphNet model (2) with these different screening-percentiles are shown in Figure 3. (c): OASIS dataset [10] with VBM. See Figure 2 for weights maps and age predictions... |

356 | Decoding mental states from brain activity in humans,” - Haynes, Rees - 2006 |

288 | Decoding the visual and subjective contents of the human brain - Kamitani, Tong - 2005 |

61 | Open access series of imaging studies (OASIS): cross-sectional MRI data in young, middle aged, nondemented, and demented older adults,”
- Marcus, Wang, et al.
- 2007
(Show Context)
Citation Context ...b): Face vs House contrast of the visual recognition dataset [9]. Weights maps obtained for the GraphNet model (2) with these different screening-percentiles are shown in Figure 3. (c): OASIS dataset =-=[10]-=- with VBM. See Figure 2 for weights maps and age predictions obtained using these different screening-percentiles. on neuro-imaging data. The first heuristic termed univariate feature-screening, provi... |

37 | Strong rules for discarding predictors in lasso-type problems.
- Tibshirani, Bien, et al.
- 2012
(Show Context)
Citation Context ...analytic expression for its proximal operator. 2i.e, techniques which don’t mistakenly discard active predictive features. those developed in [14], [15], [16], [17]. Inexact screening techniques (e.g =-=[18]-=-) have also been proposed in the literature. Our proposed heuristic screening technique is inspired by the Marginal screening technique developed in Algorithm 1 of [15], and operates as follows. The d... |

23 | Total variation regularization for fmri-based prediction of behavior
- Michel, Gramfort, et al.
- 2011
(Show Context)
Citation Context ... to perform jointly the prediction of a target variable and region segmentation in multivariate analysis settings. Specifically, it has been shown that one can employ priors like Total Variation (TV) =-=[1]-=-, TV-L1 [2], [3], TV-ElasticNet [4], and GraphNet [5] (aka SLasso [6] outside the neuroimaging community) to regularize regression and classification problems in brain imaging. The results are brain m... |

23 | Safe feature elimination in sparse supervised learning
- Wang, Ghaoui, et al.
(Show Context)
Citation Context ...ening techniques include 1That is, there is a closed-form analytic expression for its proximal operator. 2i.e, techniques which don’t mistakenly discard active predictive features. those developed in =-=[14]-=-, [15], [16], [17]. Inexact screening techniques (e.g [18]) have also been proposed in the literature. Our proposed heuristic screening technique is inspired by the Marginal screening technique develo... |

22 |
Analyses of regional-average activation and multivoxel pattern information tell complementary stories
- Jimura, Poldrack
- 2012
(Show Context)
Citation Context ...0%(|XT y|) (full-brain) respectively, survive. The green contours enclose the elite voxels which are selected by the screening procedure at the respective threshold levels. (a): Mixed Gambles dataset =-=[8]-=-. Remarkably, the geometry of the regions obtained here for the 10th and 20th screening-percentiles match pretty well the results obtained in [3] with their TV-L1 penalty. (b): Face vs House contrast ... |

18 | Lasso screening rules via dual polytope projection.
- Wang, Wonka, et al.
- 2015
(Show Context)
Citation Context ...nclude 1That is, there is a closed-form analytic expression for its proximal operator. 2i.e, techniques which don’t mistakenly discard active predictive features. those developed in [14], [15], [16], =-=[17]-=-. Inexact screening techniques (e.g [18]) have also been proposed in the literature. Our proposed heuristic screening technique is inspired by the Marginal screening technique developed in Algorithm 1... |

17 |
Interpretable whole-brain prediction analysis with graphnet,” NeuroImage,
- Grosenick, Klingenberg, et al.
- 2013
(Show Context)
Citation Context ...le and region segmentation in multivariate analysis settings. Specifically, it has been shown that one can employ priors like Total Variation (TV) [1], TV-L1 [2], [3], TV-ElasticNet [4], and GraphNet =-=[5]-=- (aka SLasso [6] outside the neuroimaging community) to regularize regression and classification problems in brain imaging. The results are brain maps which are both sparse (i.e regression coefficient... |

12 | Safe Screening with Variational Inequalities and Its Application to Lasso.
- Liu, Zhao, et al.
- 2014
(Show Context)
Citation Context ...ques include 1That is, there is a closed-form analytic expression for its proximal operator. 2i.e, techniques which don’t mistakenly discard active predictive features. those developed in [14], [15], =-=[16]-=-, [17]. Inexact screening techniques (e.g [18]) have also been proposed in the literature. Our proposed heuristic screening technique is inspired by the Marginal screening technique developed in Algor... |

8 | Structured sparsity models for brain decoding from fMRI data
- Baldassarre, Mourao-Miranda, et al.
(Show Context)
Citation Context ... jointly the prediction of a target variable and region segmentation in multivariate analysis settings. Specifically, it has been shown that one can employ priors like Total Variation (TV) [1], TV-L1 =-=[2]-=-, [3], TV-ElasticNet [4], and GraphNet [5] (aka SLasso [6] outside the neuroimaging community) to regularize regression and classification problems in brain imaging. The results are brain maps which a... |

7 | Identifying predictive regions from fMRI with TV-L1 prior
- Gramfort, Thirion, et al.
(Show Context)
Citation Context ...tly the prediction of a target variable and region segmentation in multivariate analysis settings. Specifically, it has been shown that one can employ priors like Total Variation (TV) [1], TV-L1 [2], =-=[3]-=-, TV-ElasticNet [4], and GraphNet [5] (aka SLasso [6] outside the neuroimaging community) to regularize regression and classification problems in brain imaging. The results are brain maps which are bo... |

6 | Toward open sharing of task-based fMRI data: the OpenfMRI project. Front. Neuroinform. 7:12. doi: 10.3389/fninf.2013.00012 - Poldrack, A, et al. - 2013 |

4 | Exact post model selection inference for marginal screening
- Lee, Taylor
(Show Context)
Citation Context ...techniques include 1That is, there is a closed-form analytic expression for its proximal operator. 2i.e, techniques which don’t mistakenly discard active predictive features. those developed in [14], =-=[15]-=-, [16], [17]. Inexact screening techniques (e.g [18]) have also been proposed in the literature. Our proposed heuristic screening technique is inspired by the Marginal screening technique developed in... |

3 | de Geer, “The smooth-lasso and other `1 + `2penalized methods,” Electron - Hebiri, van - 2011 |

2 |
Extracting brain regions from rest fMRI with total-variation constrained dictionary learning
- Abraham, Dohmatob, et al.
(Show Context)
Citation Context ... a structuring function (since they penalize local differences in the values of the brain map). Also, such priors produce state-of-the-art methods for automatic extraction of functional brain atlases =-=[7]-=-. However, these rich multivariate models lead to difficult optimization and model-selection problems which render them impractical on brain data. In this paper, we provide heuristic techniques for sp... |

1 | Predictive support recovery with tv-elastic net penalty and logistic regression: an application to structural mri - Dubois, Hadj-Selem, et al. |

1 |
Benchmarking solvers for tv-l1 least-squares and logistic regression in brain imaging
- Dohmatob, Gramfort, et al.
- 2014
(Show Context)
Citation Context ...nkage-Thresholding Algorithm) [12], with a provable O(1/√) convergence rate. Our implementation of FISTA uses technical recommendations (line-searching, parametrization, etc.) which were provided in =-=[13]-=-, in the context of TV-L1 [2], [3]. The model parameters α and ρ in (2) are set by internal cross-validation. (b) Univariate feature-screening: In machine-learning, feature-screening aims at detecting... |