GraphTransformer-CoGNN: A two-stage dynamic graph learning framework for label-scarce cell-type annotation in spatial transcriptomics #SingleCell 🧪🧬🖥️
https://www.researchsquare.com/article/rs-10407073/latest

A graph Transformer model predicts electronic band structures from crystal structures with DFT-level accuracy. Release: 25 Nov 2024; update: 1 Oct 2025. Read more: https://getnews.me/graph-transformer-networks-enable-accurate-band-structure-prediction/ #graphtransformer #bandstructure

HEIST, a hierarchical graph transformer, was pretrained on 22.3 million cells from 124 tissue samples across 15 organs, achieving state‑of‑the‑art results on spatial transcriptomics. Read more: https://getnews.me/heist-hierarchical-graph-transformer-for-spatial-omics/ #spatialomics #graphtransformer

G2LFormer flips the usual graph‑transformer design, using global attention early and GNN layers later, reporting higher accuracy on node classification while keeping linear cost. Read more: https://getnews.me/global-to-local-attention-improves-graph-transformer-accuracy/ #graphtransformer #gnn

GRATE fuses a Graph Transformer with a Kalman filter, reducing robot exploration travel distance by up to 21.5% and mission time by 21.3% (results 16 Sep 2025). Read more: https://getnews.me/grate-graph-transformer-boosts-robot-exploration-efficiency/ #grate #robotexploration #graphtransformer

complex graph-based tasks and applications. The source code is available in the GitHub repository at: https://github.com/xihan-qin/TSPE-GraphTransformer. [6/6 of https://arxiv.org/abs/2503.03046v1]