Advances in Graph Neural Networks and Domain Adaptation

The field of graph neural networks (GNNs) and domain adaptation is rapidly advancing, with a focus on developing more robust and generalizable models. Recent research has highlighted the importance of considering the structural properties of graphs, such as node and edge relationships, when designing GNN architectures. Additionally, there is a growing need for domain adaptation methods that can effectively transfer knowledge from one domain to another, particularly in scenarios where there is a significant shift in the distribution of data between the source and target domains. Notable papers in this area include:

  • A Benchmark Dataset for Graph Regression with Homogeneous and Multi-Relational Variants, which introduces a new benchmark dataset for graph regression tasks.
  • Target Semantics Clustering via Text Representations for Robust Universal Domain Adaptation, which proposes a novel approach for universal domain adaptation using text representations.

Sources

A Benchmark Dataset for Graph Regression with Homogeneous and Multi-Relational Variants

Bridging Source and Target Domains via Link Prediction for Unsupervised Domain Adaptation on Graphs

Conformal Prediction for Zero-Shot Models

From Features to Structure: Task-Aware Graph Construction for Relational and Tabular Learning with GNNs

Theoretical Performance Guarantees for Partial Domain Adaptation via Partial Optimal Transport

Target Semantics Clustering via Text Representations for Robust Universal Domain Adaptation

Out-of-Distribution Graph Models Merging

How to Use Graph Data in the Wild to Help Graph Anomaly Detection?

OpenGT: A Comprehensive Benchmark For Graph Transformers

iN2V: Bringing Transductive Node Embeddings to Inductive Graphs

Built with on top of