Holographic Neural Architectures

TL;DR

Holographic Neural Architectures (HNAs) introduce a novel representation learning framework inspired by holography, enabling robust, noise-resistant models that excel in generative tasks and are suitable for biological data with limited or noisy training samples.

Contribution

This work presents HNAs, a new framework for representation learning that produces holographic representations, enhancing generative capabilities and noise resistance in neural models.

Findings

HNAs enable effective generative modeling.

HNAs exhibit high resistance to noise.

HNAs are well-suited for biological applications with limited data.

Abstract

Representation learning is at the heart of what makes deep learning effective. In this work, we introduce a new framework for representation learning that we call "Holographic Neural Architectures" (HNAs). In the same way that an observer can experience the 3D structure of a holographed object by looking at its hologram from several angles, HNAs derive Holographic Representations from the training set. These representations can then be explored by moving along a continuous bounded single dimension. We show that HNAs can be used to make generative networks, state-of-the-art regression models and that they are inherently highly resistant to noise. Finally, we argue that because of their denoising abilities and their capacity to generalize well from very few examples, models based upon HNAs are particularly well suited for biological applications where training examples are rare or noisy.