R-Mixup: Riemannian Mixup for Biological Networks

Abstract

Biological networks are commonly used in biomedical and healthcare domains to effectively model the structure of complex biological systems with interactions linking biological entities. However, due to their characteristics of high dimensionality and low sample size, directly applying deep learning models on biological networks usually faces severe overfitting. In this work, we propose R-MIXUP, a Mixup-based data augmentation technique that suits the symmetric positive definite (SPD) property of adjacency matrices from biological networks with optimized training efficiency. The interpolation process in R-MIXUP leverages the log-Euclidean distance metrics from the Riemannian manifold, effectively addressing the swelling effect and arbitrarily incorrect label issues of vanilla Mixup. We demonstrate the effectiveness of R-MIXUP with five real-world biological network datasets on both regression and classification tasks. Besides, we derive a commonly ignored necessary condition for identifying the SPD matrices of biological networks and empirically study its influence on the model performance. The code implementation can be found in Appendix E.

Accepted to KDD 2023

References in corpus (22)

• Twins: Revisiting the Design of Spatial Attention in Vision Transformers
• Non-Euclidean statistics for covariance matrices, with applications to diffusion tensor imaging
• Kernel Methods on Riemannian Manifolds with Gaussian RBF Kernels
• BrainGB: A Benchmark for Brain Network Analysis with Graph Neural Networks
• Local polynomial regression on unknown manifolds
• Brain Network Transformer
• G-Mixup: Graph Data Augmentation for Graph Classification
• FBNETGEN: Task-aware GNN-based fMRI Analysis via Functional Brain Network Generation
• MAtt: A Manifold Attention Network for EEG Decoding
• Data-Efficient Brain Connectome Analysis via Multi-Task Meta-Learning
• Weakly-Supervised Scientific Document Classification via Retrieval-Augmented Multi-Stage Training
• C-Mixup: Improving Generalization in Regression
• GraphMixup: Improving Class-Imbalanced Node Classification on Graphs by Self-supervised Context Prediction
• Cold-Start Data Selection for Few-shot Language Model Fine-tuning: A Prompt-Based Uncertainty Propagation Approach
• Deep Representations for Time-varying Brain Datasets
• Learning Task-Aware Effective Brain Connectivity for fMRI Analysis with Graph Neural Networks
• Multi-objective Deep Data Generation with Correlated Property Control
• PTGB: Pre-Train Graph Neural Networks for Brain Network Analysis
• Local Boosting for Weakly-Supervised Learning
• BrainNet: Epileptic Wave Detection from SEEG with Hierarchical Graph Diffusion Learning
• A Probabilistic Bayesian Approach to Recover $R_2^{\ast }$ map and Phase Images for Quantitative Susceptibility Mapping
• Domain Generalization Deep Graph Transformation