Graph Element Networks: adaptive, structured computation and memory

Authors: Ferran Alet, Adarsh Keshav Jeewajee, Maria Bauza Villalonga, Alberto Rodriguez, Tomas Lozano-Perez, Leslie Kaelbling

ICML 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We do a varied set of experiments to illustrate the diversity of applications of GENs.
Researcher Affiliation Academia 1CSAIL MIT, Cambridge, MA, USA 2Mechanical Engineering MIT, Cambridge, MA, USA.
Pseudocode No The paper mentions an "algorithm for training it from data" (Section 1) but does not present any structured pseudocode or algorithm blocks.
Open Source Code Yes Code can be found at https://github.com/FerranAlet/graph_element_networks.
Open Datasets Yes For training, we get ground-truth solutions using a FEM with a very dense mesh of 2502 nodes, computed with the program FEni CS (Alnæs et al., 2015).
Dataset Splits No The paper mentions training and evaluation on different sets, such as "trained on a set of houses ... and evaluated on a different set" (Section 4.1) and "measured on held-out data" (Section 4.2), but it does not provide specific percentages, absolute counts, or detailed methodologies for the train/validation/test splits in the main text.
Hardware Specification No The paper does not provide specific hardware details such as GPU or CPU models, or cloud computing instance types used for running the experiments.
Software Dependencies No The paper mentions software like "FEniCS" (Section 4.1) but does not provide specific version numbers for any software components or libraries used in the experiments.
Experiment Setup Yes We use a set of k2 nodes placed on a uniformly-spaced grid, and a number of message passing steps equal to the diameter of the graph, T = 2(k - 1).