A Local-Ascending-Global Learning Strategy for Brain-Computer Interface

Authors: Dongrui Gao, Haokai Zhang, Pengrui Li, Tian Tang, Shihong Liu, Zhihong Zhou, Shaofei Ying, Ye Zhu, Yongqing Zhang

AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The proposed LAG strategy is validated using datasets related to fatigue (SEED-VIG), emotion (SEED-IV), and motor imagery (BCI C IV 2a). The results demonstrate the generalizability of LAG, achieving satisfactory outcomes in independent-subject experiments across all three datasets.
Researcher Affiliation Academia 1School of Computer Science, Chengdu University of Information Technology, Chengdu, 610225, China 2 School of Life Sciences and Technology, University of Electronic Science and Technology of China, Chengdu, 611731, China
Pseudocode Yes Algorithm 1: Training Stage Input: Train set G = {BM} Output: {ypred, Loss}
Open Source Code No The paper does not provide any explicit statement or link for open-source code availability for the described methodology.
Open Datasets Yes This paper presents validation experiments conducted on three EEG datasets (SEED-VIG, SEED-IV, BCI C IV 2a) each associated with distinct cognitive tasks (Gao et al. 2023a; Peng et al. 2023; Zhang et al. 2019).
Dataset Splits No The paper mentions 'independent-subject experiments' and 'cross-subject comparison results' but does not specify exact split percentages or sample counts for training, validation, or test sets.
Hardware Specification No The paper does not provide specific details about the hardware (e.g., GPU/CPU models, memory) used for running the experiments.
Software Dependencies No The paper mentions the 'Adam optimizer' but does not provide specific version numbers for any software dependencies or frameworks used.
Experiment Setup Yes The graph convolution order is set to 2, and a dropout rate of 0.5 is applied. Model parameters are optimized using the Adam optimizer, with a learning rate search range of [1e-3, 1e-1] and an L2 regularization search range of [5e-3, 3e-1].