invariance

Geom3D: Geometric Modeling on 3D Structures, NeurIPS 2023

Equivariant Subgraph Aggregation Networks (ICLR 2022 Spotlight)

Official PyTorch and JAX Implementation of "Harmonics of Learning: Universal Fourier Features Emerge in Invariant Networks"

Size-Invariant Graph Representations for Graph Classification Extrapolations (ICML 2021 Long Talk)

Continuous regular group convolutions for Pytorch

On the forward invariance of Neural ODEs: performance guarantees for policy learning

Code for "Improving Stain Invariance of CNNs for Segmentation by Fusing Channel Attention and Domain-Adversarial Training"

Official PyTorch Implementation of "A General Framework for Robust G-Invariance in G-Equivariant Networks," NeurIPS 2023

Official code for NeurIPS 2024 paper "A Canonicalization Perspective on Invariant and Equivariant Learning".

(EAAI) Source-code of the paper: Radon averaging: A practical approach for designing rotation-invariant models

The Transformational Measures (TM) library allows neural network researchers to evaluate the invariance and equivariance of their models with respect to a set of transformations. Support for Pytorch (current) and Tensorflow/Keras (coming).

Implementation for the NeurIPS 2025 paper: An Analysis of Causal Effect Estimation using Outcome Invariant Data Augmentation

Official implementation of the paper "Designing Affine-Invariant NN For Photometric Corruption Robustness and Generalization" (ICLR'26)

Official code for NeurIPS 2024 paper "A Canonicalization Perspective on Invariant and Equivariant Learning".

Code for the Obj-MNIST Dataset generation framework. Proposed in "Parameter-Efficient Invariance via Equivariant Object Detection Networks".

CL-DIAG: Diagnosing canonicalization leakage in supervised learning under group symmetry

STELA — a dependency-free symbolic grammar of meaning (9 invariants + 5 anchors as primes), with the whole Bible (all 66 books, 90 stations) translated, invariance-proven, and rendered to audio.

Identifying invariant elements of Neural Networks to image transformations

Stress-test suite for agent output invariance and reliability — by Steven Crawford-Maggard (EVEZ)

The is a forensic auditing system designed to detect, measure, and document systematic degradation of technical truth in corporate AI models. Through rigorous application of information theory, thermodynamic principles, and cryptographic sovereignty, quantifies censorship.