DiffLifting

This repository is the official implementation of the methods in the publication Differentiable Lifting for Topological Neural Networks, accepted at ICLR 2026. The code runs on python 3.10 and uses CUDA-12.1.

Description

We propose $\partial\text{lift}$ (DiffLift), a general framework for learning graph liftings to hypergraphs and cellular- and simplicial complexes in an end-to-end fashion. In particular, our approach leverages learned vertex-level latent representations to identify and parameterize distributions over candidate higher-order cells for inclusion. This results in a scalable model which can be readily integrated into any TNN. Our experiments show that $\partial\text{lift}$ outperforms existing lifting methods on multiple benchmarks for graph and node classification across different TNN architectures. Notably, our approach leads to gains of up to 45% over static liftings, including both connectivity- and feature-based ones.

How to run

docker build .

Experiments

Graph classification experiments are in main_graph_classification.py, and node classification experiments are in main.py.

Citation

If you find this code useful in your research, please consider citing our paper:

@inproceedings{
franco2026differentiable,
title={Differentiable Lifting for Topological Neural Networks},
author={Jorge Luiz Franco and Gabriel Duarte and Alexander V Nikitin and Moacir A Ponti and Diego Mesquita and Amauri H Souza},
booktitle={The Fourteenth International Conference on Learning Representations},
year={2026},
url={[https://openreview.net/forum?id=eC89CbINIw](https://openreview.net/forum?id=eC89CbINIw)}
}

License

This software is provided under the MIT License.