Homological Neural Networks: A Sparse Architecture for Multivariate Complexity
This addresses efficiency and interpretability challenges for AI practitioners, though it appears incremental as it builds on existing network-based filtering techniques.
The paper tackles computational complexity and interpretability in deep learning by introducing a sparse higher-order graphical neural network unit based on homological data structures, achieving results that match or exceed state-of-the-art models in tabular data and time series regression with significantly fewer parameters.
The rapid progress of Artificial Intelligence research came with the development of increasingly complex deep learning models, leading to growing challenges in terms of computational complexity, energy efficiency and interpretability. In this study, we apply advanced network-based information filtering techniques to design a novel deep neural network unit characterized by a sparse higher-order graphical architecture built over the homological structure of underlying data. We demonstrate its effectiveness in two application domains which are traditionally challenging for deep learning: tabular data and time series regression problems. Results demonstrate the advantages of this novel design which can tie or overcome the results of state-of-the-art machine learning and deep learning models using only a fraction of parameters.