gnn

Proposes a dual-primal graph VAE architecture that treats ground-truth labels as latent variables to aggregate noisy crowdsourced labels, achieving state-of-the-art results on crowdsourcing benchmarks without needing a separate classifier.

Presents SNI-GNN, a SmartNIC-assisted full-graph GNN training system that reduces inter-node communication by predicting remote embeddings in-network, achieving 1.3–3.6x speedups with negligible accuracy loss.

A computer science paper presents a graph-based workflow serving engine that unifies agent operations into a global wGraph, using dynamic graph synthesis and differential KV-cache reuse to boost agent accuracy by 4.95% while cutting GPU memory usage by 4x.

A curated collection of GNN papers, datasets, and implementation tools, hosted on GitHub.

This paper proposes LLM-GNN Co-Teaching, a bidirectional framework for few-shot graph learning on text-attributed graphs. The LLM and GNN exchange confident pseudo-labels and use round-based preference optimization (RPL-PO) to mutually improve, outperforming prior methods on benchmarks.

This paper proposes a probabilistic contrastive pretraining framework for molecular graph transformers to improve multi-task ADME property prediction in drug discovery, achieving significant gains on three benchmarks.

This survey examines computational nondeterminism in financial AI systems, covering tabular models, graph networks, and LLM-based workflows, and proposes a layered evaluation framework for auditability.