On Universal Equivariant Set Networks

Authors: Nimrod Segol, Yaron Lipman

ICLR 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Lastly, we provide numerical experiments validating the theoretical results and comparing different permutation equivariant models.
Researcher Affiliation Academia Nimrod Segol & Yaron Lipman Department of Computer Science and Applied Mathematics Weizmann Institute of Science Rehovot, Israel {nimrod.segol,yaron.lipman}@weizmann.ac.il
Pseudocode No The paper does not contain structured pseudocode or algorithm blocks.
Open Source Code Yes The code can be found at https://github.com/Nimrod Segol/On-Universal-Equivariant-Set-Networks
Open Datasets Yes We used the Model Net dataset (Wu et al., 2015)
Dataset Splits Yes We drew 10k training examples and 1k test examples i.i.d. from a N( 1/2, 1) distribution (per entry of X).
Hardware Specification No The paper does not provide specific hardware details (exact GPU/CPU models, processor types, or memory amounts) used for running its experiments.
Software Dependencies Yes We implemented the experiments in Pytorch Paszke et al. (2017) with the Adam Kingma & Ba (2014) optimizer for learning.
Experiment Setup Yes For the classification we used the cross entropy loss and trained for 150 epochs with learning rate 0.001, learning rate decay of 0.5 every 100 epochs and batch size 32. For the quadratic function regression we trained for 150 epochs with leaning rate of 0.001, learning rate decay 0.1 every 50 epochs and batch size 64; for the regression to the leading eigen vector we trained for 50 epochs with leaning rate of 0.001 and batch size 32. To regress to the output of a single graph convolution layer we trained for 200 epochs with leaning rate of 0.001 and batch size 32.