Learning the Morphology of Brain Signals Using Alpha-Stable Convolutional Sparse Coding
Authors: Mainak Jas, Tom Dupré la Tour, Umut Simsekli, Alexandre Gramfort
NeurIPS 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We illustrate the benefits of the proposed approach on both synthetic and real datasets. In order to evaluate our approach, we conduct several experiments on both synthetic and real data. |
| Researcher Affiliation | Academia | Mainak Jas1, Tom Dupré La Tour1, Umut Sim sekli1, Alexandre Gramfort1,2 1: LTCI, Telecom Paris Tech, Université Paris-Saclay, Paris, France 2: INRIA, Université Paris-Saclay, Saclay, France |
| Pseudocode | Yes | Algorithm 1 α-stable Convolutional Sparse Coding |
| Open Source Code | Yes | The source code is publicly available at https://alphacsc.github.io/. |
| Open Datasets | Yes | We first applied αCSC on an LFP dataset previously used in [8] and containing epileptiform spikes as shown in Fig. 4(a). The second dataset is an LFP channel in a rodent striatum from [35]. |
| Dataset Splits | No | The paper uses synthetic and real datasets ('N trials of length T', '70 trials of length 2500 samples') but does not explicitly describe standard training, validation, and test splits (e.g., '80/10/10 split') for model evaluation. The focus is on learning atoms from the data itself and evaluating the robustness of this learning under various conditions, such as data corruption. |
| Hardware Specification | No | The paper does not provide specific details regarding the hardware used for running experiments, such as GPU models, CPU types, or memory specifications. It mentions 'single-threaded and a parallel version for the z-update' but without hardware context. |
| Software Dependencies | No | The paper does not specify version numbers for any software dependencies or libraries used in the implementation or experiments. It refers to algorithms like L-BFGS-B and ADMM-based methods, but without version details. |
| Experiment Setup | Yes | In our first set of synthetic experiments, we set N = 100, T = 2000 and λ = 1, and use different values for K and L. To be comparable, we set α = 2 and add Gaussian noise to the synthesized signals, where the standard deviation is set to 0.01. [...] For αCSC, we set the number of outer iterations I = 5, the number of iterations of the M-step to M = 50, and the number of iterations of the MCMC algorithm to J = 10. We discard the first 5 samples of the MCMC algorithm as burn-in. |