Lie Neurons: Adjoint-Equivariant Neural Networks for Semisimple Lie Algebras

Authors: Tzu-Yuan Lin, Minghan Zhu, Maani Ghaffari

ICML 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments are conducted for the so(3), sl(3), and sp(4) Lie algebras on various tasks, including fitting equivariant and invariant functions, learning system dynamics, point cloud registration, and homography-based shape classification.
Researcher Affiliation Academia 1University of Michigan, Ann Arbor, MI, USA.
Pseudocode No The paper describes methods in text and figures (e.g., Figure 5 shows architecture diagrams) but does not contain a formal 'Pseudocode' or 'Algorithm' block.
Open Source Code Yes The software implementation is available at https://github.com/UMich-CURLY/Lie Neurons.
Open Datasets Yes The network is trained and evaluated on Model Net40, which contains 3D models of objects in 40 categories.
Dataset Splits No The paper consistently mentions 'training' and 'testing' data, but does not explicitly provide details about a separate 'validation' dataset split for reproduction.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory) used for running its experiments.
Software Dependencies No The paper does not explicitly list specific software dependencies with version numbers. It refers readers to the GitHub repository for implementation details.
Experiment Setup Yes The network architecture used in this experiment can be found in Figure 5 in the Appendix. It consists of two LN-LB+LN-LR layers. The feature dimension of each layer is set to 1024, while the last linear layer projects the features back to dimension 3.