Latent Graph Inference

Latent graph inference (LGI) focuses on learning the underlying relational structure between data points when the explicit graph is unavailable or noisy, enabling more effective analysis using graph neural networks (GNNs). Current research emphasizes developing improved GNN architectures that jointly learn node representations and graph structures, often incorporating advanced techniques like Boolean product operations, probability passing, and attention mechanisms to refine graph inference and improve robustness. This field is significant because it allows GNNs to be applied to a wider range of problems where relational information is implicit, impacting diverse applications from time series forecasting to point cloud analysis. Furthermore, research is exploring optimal embedding spaces (e.g., hyperbolic spaces) and efficient algorithms to address computational challenges associated with large datasets.

Papers

September 21, 2024

Boolean Product Graph Neural Networks
Ziyan Wang, Bin Liu, Ling Xiang
Graph Neural Network Latent Graph Boolean Network Latent Graph Inference Graph Product

July 15, 2024

Probability Passing for Graph Neural Networks: Graph Structure and Representations Joint Learning
Ziyan Wang, Yaxuan He, Bin Liu
Graph Neural Network Graph Structure Graph Data Joint Representation Probability Pr(E1 Latent Graph Inference

November 20, 2023

AMES: A Differentiable Embedding Space Selection Framework for Latent Graph Inference
Yuan Lu, Haitz Sáez de Ocáriz Borde, Pietro Liò
Graph Neural Network Closed Form Differentiable Expression Latent Graph Latent Graph Inference

October 23, 2023

Neural Snowflakes: Universal Latent Graph Inference via Trainable Latent Geometries
Haitz Sáez de Ocáriz Borde, Anastasis Kratsios
Latent Graph Latent Geometry Latent Graph Inference Neural Snowflake

October 6, 2023

Latent Graph Inference with Limited Supervision
Jianglin Lu, Yi Xu, Huan Wang, Yue Bai, Yun Fu
Node Representation Partial Supervision Graph Sparsification Edge Weight Latent Graph Inference Weak Supervision Signal

May 25, 2023

From Latent Graph to Latent Topology Inference: Differentiable Cell Complex Module
Claudio Battiloro, Indro Spinelli, Lev Telyatnikov, Michael Bronstein, Simone Scardapane, Paolo Di Lorenzo
Graph Neural Network Latent Graph Latent Topology Topological Deep Learning Latent Graph Inference

March 21, 2023

Projections of Model Spaces for Latent Graph Inference
Haitz Sáez de Ocáriz Borde, Álvaro Arroyo, Ingmar Posner
Graph Neural Network Projection Bias Camera Projection Model Space Latent Graph Inference

February 14, 2023

Graph schemas as abstractions for transfer learning, inference, and planning
J. Swaroop Guntupalli, Rajkumar Vasudeva Raju, Shrinu Kushagra, Carter Wendelken, Danny Sawyer, Ishan Deshpande, Guangyao Zhou, Miguel Lázaro-Gredilla, Dileep George
Transfer Learning Scientific Inference Latent Structure Latent Graph Metric Aware Abstraction Latent Graph Inference Schema Graph

November 26, 2022

Latent Graph Inference using Product Manifolds
Haitz Sáez de Ocáriz Borde, Anees Kazi, Federico Barbero, Pietro Liò
Latent Graph Product Manifold Latent Graph Inference

November 7, 2022

pyGSL: A Graph Structure Learning Toolkit
Max Wasserman, Gonzalo Mateos
Graph Structure Graph Structure Learning Latent Graph Inference Differentiable Graph

March 7, 2022

Multivariate Time Series Forecasting with Latent Graph Inference
Victor Garcia Satorras, Syama Sundar Rangapuram, Tim Januschowski
Multivariate Time Series Forecasting Forecasting Performance Graph Inference Latent Graph Inference