Capacity and Bias of Learned Geometric Embeddings for Directed Graphs

Authors: Michael Boratko, Dongxu Zhang, Nicholas Monath, Luke Vilnis, Kenneth L Clarkson, Andrew McCallum

NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We perform rigorous empirical evaluations of vector, hyperbolic, and region-based geometric representations on several families of synthetic and realworld directed graphs.
Researcher Affiliation Collaboration 1 University of Massachusetts Amherst 2 IBM Research
Pseudocode No The paper describes methods and models using mathematical equations and textual explanations but does not include structured pseudocode or algorithm blocks.
Open Source Code Yes Our code and data are available at https://github.com/iesl/geometric_graph_embedding.
Open Datasets Yes Graphs Real World. We also select the following real world graph datasets: Word Net (Animals) [34]... Hierarchical Clustering. We run agglomerative clustering on the Inception V3 [56] features from 213 Image Net images [47].
Dataset Splits No We check the training loss ten times per epoch, and apply early-stopping with a patience of just over 2 epochs (21 loss observations). The paper mentions early stopping which implies a validation set, but it does not specify explicit training/validation/test splits (e.g., percentages or sample counts) for the datasets used.
Hardware Specification No The paper mentions 'high performance computing equipment' in the acknowledgements but does not provide specific hardware details such as exact GPU/CPU models or memory amounts used for experiments.
Software Dependencies No The paper mentions using W&B [5] for hyperparameter optimization but does not provide specific version numbers for software dependencies or libraries used in the implementation.
Experiment Setup Yes All models are tuned on learning rate, batch size, and weight of negative loss. We tune the margin γ for OE, and β parameters in (9) for HYPERBOLIC, intersection and volume temperature for BOX, and the initialization of these temperatures for T-BOX. We check the training loss ten times per epoch, and apply early-stopping with a patience of just over 2 epochs (21 loss observations).