| 1. |
|
|
| 2. |
|
|
| 3. |
|
|
| 4. |
- Moreau, Thomas, et al.
(författare)
-
Benchopt : Reproducible, efficient and collaborative optimization benchmarks
- 2022
-
Ingår i: Advances in Neural Information Processing Systems 35 - 36th Conference on Neural Information Processing Systems, NeurIPS 2022. - 1049-5258. - 9781713871088 ; 35, s. 25404-25421
-
Konferensbidrag (refereegranskat)abstract
- Numerical validation is at the core of machine learning research as it allows to assess the actual impact of new methods, and to confirm the agreement between theory and practice. Yet, the rapid development of the field poses several challenges: researchers are confronted with a profusion of methods to compare, limited transparency and consensus on best practices, as well as tedious re-implementation work. As a result, validation is often very partial, which can lead to wrong conclusions that slow down the progress of research. We propose Benchopt, a collaborative framework to automate, reproduce and publish optimization benchmarks in machine learning across programming languages and hardware architectures. Benchopt simplifies benchmarking for the community by providing an off-the-shelf tool for running, sharing and extending experiments. To demonstrate its broad usability, we showcase benchmarks on three standard learning tasks: ℓ2-regularized logistic regression, Lasso, and ResNet18 training for image classification. These benchmarks highlight key practical findings that give a more nuanced view of the state-of-the-art for these problems, showing that for practical evaluation, the devil is in the details. We hope that Benchopt will foster collaborative work in the community hence improving the reproducibility of research findings.
|
|
| 5. |
- Niso, G, et al.
(författare)
-
MEG-BIDS, the brain imaging data structure extended to magnetoencephalography
- 2018
-
Ingår i: Scientific data. - : Springer Science and Business Media LLC. - 2052-4463. ; 5, s. 180110-
-
Tidskriftsartikel (refereegranskat)abstract
- We present a significant extension of the Brain Imaging Data Structure (BIDS) to support the specific aspects of magnetoencephalography (MEG) data. MEG measures brain activity with millisecond temporal resolution and unique source imaging capabilities. So far, BIDS was a solution to organise magnetic resonance imaging (MRI) data. The nature and acquisition parameters of MRI and MEG data are strongly dissimilar. Although there is no standard data format for MEG, we propose MEG-BIDS as a principled solution to store, organise, process and share the multidimensional data volumes produced by the modality. The standard also includes well-defined metadata, to facilitate future data harmonisation and sharing efforts. This responds to unmet needs from the multimodal neuroimaging community and paves the way to further integration of other techniques in electrophysiology. MEG-BIDS builds on MRI-BIDS, extending BIDS to a multimodal data structure. We feature several data-analytics software that have adopted MEG-BIDS, and a diverse sample of open MEG-BIDS data resources available to everyone.
|
|
| 6. |
|
|
