Graph neural networks are inherently good generalizers: Insights by bridging gnns and mlps

C Yang, Q Wu, J Wang, J Yan - arxiv preprint arxiv:2212.09034, 2022 - arxiv.org
Graph neural networks (GNNs), as the de-facto model class for representation learning on
graphs, are built upon the multi-layer perceptrons (MLP) architecture with additional …

Graph Machine Learning through the Lens of Bilevel Optimization

AY Zheng, T He, Y Qiu, M Wang… - … Conference on Artificial …, 2024 - proceedings.mlr.press
Bilevel optimization refers to scenarios whereby the optimal solution of a lower-level energy
function serves as input features to an upper-level objective of interest. These optimal …

BloomGML: Graph Machine Learning through the Lens of Bilevel Optimization

AY Zheng, T He, Y Qiu, M Wang, D Wipf - arxiv preprint arxiv:2403.04763, 2024 - arxiv.org
Bilevel optimization refers to scenarios whereby the optimal solution of a lower-level energy
function serves as input features to an upper-level objective of interest. These optimal …

Beyond Graph Convolution: Multimodal Recommendation with Topology-aware MLPs

J Huang, J Qin, Y Yu, W Zhang - arxiv preprint arxiv:2412.11747, 2024 - arxiv.org
Given the large volume of side information from different modalities, multimodal
recommender systems have become increasingly vital, as they exploit richer semantic …

Preventing Model Collapse in Deep Canonical Correlation Analysis by Noise Regularization

J He, J Du, S Xu, W Ma - arxiv preprint arxiv:2411.00383, 2024 - arxiv.org
Multi-View Representation Learning (MVRL) aims to learn a unified representation of an
object from multi-view data. Deep Canonical Correlation Analysis (DCCA) and its variants …