Multivariate Convolutional Sparse Coding for Electromagnetic Brain Signals
Authors: Tom Dupré la Tour, Thomas Moreau, Mainak Jas, Alexandre Gramfort
NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | All numerical experiments were run using Python (Python Software Foundation, 2017) and our code is publicly available online at https://alphacsc.github.io/. |
| Researcher Affiliation | Academia | 1: LTCI, Télécom Paris Tech, Université Paris-Saclay, Paris, France 2: INRIA, Université Paris-Saclay, Saclay, France |
| Pseudocode | Yes | Algorithm 1: Locally greedy coordinate descent (LGCD) |
| Open Source Code | Yes | All numerical experiments were run using Python (Python Software Foundation, 2017) and our code is publicly available online at https://alphacsc.github.io/. |
| Open Datasets | Yes | The somatosensory dataset from the MNE software (Gramfort et al., 2013, 2014) contains responses to median nerve stimulation. |
| Dataset Splits | No | The paper describes the characteristics of the data used for experiments but does not specify explicit training, validation, or test splits, or a methodology for creating them for reproducibility beyond general parameters like N, K, L. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., GPU/CPU models, memory amounts, or detailed computer specifications) used for running its experiments. |
| Software Dependencies | No | The paper mentions "Python (Python Software Foundation, 2017)" which implies Python 3.6, but does not list other key software components or libraries with specific version numbers (e.g., MNE software is mentioned but without a version). |
| Experiment Setup | Yes | First we compared our strategy against three state-of-the-art univariate CSC solvers available online. The first was developed by Garcia-Cardona and Wohlberg (2017) and is based on ADMM. The second and third were developed by Jas et al. (2017), and are respectively based on FISTA and L-BFGS. All solvers shared the same objective function, but as the problem is non-convex, the solvers are not guaranteed to reach the same local minima, even though we started from the same initial settings. Hence, for a fair comparison, we computed the convergence curves relative to each local minimum, and averaged them over 10 different initializations. The somatosensory dataset from the MNE software (Gramfort et al., 2013, 2014) contains responses to median nerve stimulation. We consider only gradiometers channels and we used the following parameters: T = 134 700, N = 2, K = 8, and L = 128. We learned K = 40 atoms with L = 150 using a rank-1 multivariate CSC model, with a regularization λ = 0.2λmax. |