Adapting graph neural network connectivity and resolution to the noise level improves flow-based generation of 3D point clouds, spatial transcriptomics, and images.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Geometric Generative Modeling with Noise-Conditioned Graph Networks
Adapting graph neural network connectivity and resolution to the noise level improves flow-based generation of 3D point clouds, spatial transcriptomics, and images.