Spectra defines and controls effective capacity in graph embeddings via the Shannon effective rank of a trace-normalized kernel spectrum, making capacity a post-fit property rather than a pre-training hyperparameter.
What do gnns actually learn? towards understanding their representations
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 3years
2026 3verdicts
UNVERDICTED 3roles
background 1polarities
background 1representative citing papers
Graph nodes are embedded as simplex compositions via ILR coordinates in Aitchison geometry to obtain interpretable representations that support component restriction and competitive task performance.
TACENR introduces a contrastive-learning method that identifies the most influential attribute, proximity, and structural features in node representations in a task-agnostic manner.
citing papers explorer
-
Rank Is Not Capacity: Spectral Occupancy for Latent Graph Models
Spectra defines and controls effective capacity in graph embeddings via the Shannon effective rank of a trace-normalized kernel spectrum, making capacity a post-fit property rather than a pre-training hyperparameter.
-
Aitchison Embeddings for Learning Compositional Graph Representations
Graph nodes are embedded as simplex compositions via ILR coordinates in Aitchison geometry to obtain interpretable representations that support component restriction and competitive task performance.
-
TACENR: Task-Agnostic Contrastive Explanations for Node Representations
TACENR introduces a contrastive-learning method that identifies the most influential attribute, proximity, and structural features in node representations in a task-agnostic manner.