Granular computing with multiple granular layers for brain big data processing

Decoupling Representation Learning for Imbalanced Electroencephalography Classification in Rapid Serial Visual Presentation Task

OBJECTIVE

The class imbalance problem considerably restricts the performance of electroencephalography (EEG) classification in the rapid serial visual presentation (RSVP) task. Existing solutions typically employ re-balancing strategies (e.g. re-weighting and re-sampling) to alleviate the impact of class imbalance, which enhances the classifier learning of deep networks but unexpectedly damages the representative ability of the learned deep features as original distributions become distorted.

APPROACH

In this study, a novel decoupling representation learning (DRL) model, has been proposed that separates the representation learning and classification processes to capture the discriminative feature of imbalanced RSVP EEG data while classifying it accurately. The representation learning process is responsible for learning universal patterns for the classification of all samples, while the classifier determines a better bounding for the target and non-target classes. Specifically, the representation learning process adopts a dual-branch architecture, which minimizes the contrastive loss to regularize the representation space. In addition, to learn more discriminative information from RSVP EEG data, a novel multi-granular information (MGI) based extractor is designed to extract spatial-temporal information. Considering the class re-balancing strategies can significantly promote classifier learning, the classifier was trained with rebalanced EEG data while freezing the parameters of the representation learning process.

MAIN RESULTS

To evaluate the proposed method, experiments were conducted on two public datasets and one self-conducted dataset. The results demonstrate that the proposed DRL can achieve state-of-the-art performance for EEG classification in the RSVP task.

SIGNIFICANCE

This is the first study to focus on the class imbalance problem and propose a generic solution in the RSVP task. Furthermore, multi-granular data was explored to extract more complementary spatial-temporal information. The code is open-source and available at https://github.com/Tammie-Li/DRL.

Multi-granularity Complex Network Representation Learning

Network representation learning aims to learn the low dimensional vector of the nodes in a network while maintaining the inherent properties of the original information. Existing algorithms focus on the single coarse-grained topology of nodes or text information alone, which cannot describe complex information networks. However, node structure and attribution are interdependent, indecomposable. Therefore, it is essential to learn the representation of node based on both the topological structure and node additional attributes. In this paper, we propose a multi-granularity complex network representation learning model (MNRL), which integrates topological structure and additional information at the same time, and presents these fused information learning into the same granularity semantic space that through fine-to-coarse to refine the complex network. Experiments show that our method can not only capture indecomposable multi-granularity information, but also retain various potential similarities of both topology and node attributes. It has achieved effective results in the downstream work of node classification and the link prediction on real-world datasets.