# Modality-Agnostic Variational Compression of Implicit Neural Representations

**Authors:** Jonathan Richard Schwarz, Jihoon Tack, Yee Whye Teh, Jaeho Lee, Jinwoo Shin  
**arXiv:** [2301.09479](https://arxiv.org/abs/2301.09479)  
**Date:** 2023-01-23  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/2301.09479v3/

## Key points

- Authors introduce a modality-agnostic neural compression algorithm based on Implicit Neural Representations (INR)
- Algorithm addresses gap between latent coding and sparsity by generating compact latent representations mapped to soft gating mechanism
- Specialization of shared INR network to individual data items through subnetwork selection improves rate/distortion trade-off in modality-agnostic space
- Proposed approach, Variational Compression of INRs (VC-INR), outperforms previous quantization schemes for INR techniques
- VC-INR demonstrates superior performance compared to well-known codecs like JPEG 2000, MP3, and AVC/HEVC in various modalities
- Research leverages INRs and variational methods to achieve state-of-the-art results across diverse data types without modality-specific adjustments

## Summary

In their paper titled "Modality-Agnostic Variational Compression of Implicit Neural Representations," authors Jonathan Richard Schwarz, Jihoon Tack, Yee Whye Teh, Jaeho Lee, and Jinwoo Shin introduce a novel neural compression algorithm that is modality-agnostic and based on a functional perspective of data. This algorithm is parameterized as an Implicit Neural Representation (INR) and addresses the gap between latent coding and sparsity by generating compact latent representations that are non-linearly mapped to a soft gating mechanism. By allowing for the specialization of a shared INR network to individual data items through subnetwork selection, the authors achieve improved performance in optimizing the rate/distortion trade-off in a modality-agnostic space using neural compression. The proposed approach, known as Variational Compression of Implicit Neural Representations (VC-INR), demonstrates superior performance compared to previous quantization schemes used for other INR techniques. Through experiments across various modalities such as images, climate data, 3D shapes and scenes, audio, and video, the authors showcase the effectiveness of VC-INR without relying on modality-specific biases. Notably, VC-INR surpasses well-known codecs like JPEG 2000, MP3, and AVC/HEVC in their respective modalities. Overall,this research presents a significant advancement in neural compression techniques by leveraging INRs and variational methods to achieve state-of-the-art results across diverse data types without the need for modality-specific adjustments.