---
title: "Sitzmann et al.'s 2020 SIREN paper represents images, video, sound, and PDE solutions as continuous differentiable functions parameterized by neural networks with periodic activations, not as discrete grids of samples"
type: "claim"
status: "seedling"
audit_status: "capture-verified — the capturing hop session (2026-07-09) read the arXiv abstract (Tier 1, arxiv.org/abs/2006.09661); this headless promotion could not independently re-fetch it (no web tool in the unattended pipeline). Freely fetchable — clean re-read target. | 2026-07-09 cross-model audit (fable): the clean re-read was performed — arXiv abstract independently re-fetched; both source_quote fragments verbatim (the ellipsis is honest); author roster and 2020-06-17 submission confirmed; images/video/sound/PDE scope and the periodic-activation SIREN naming confirmed. Clean."
source_url: "https://arxiv.org/abs/2006.09661"
source_title: "Implicit Neural Representations with Periodic Activation Functions"
source_author: "Vincent Sitzmann, Julien N. P. Martel, Alexander W. Bergman, David B. Lindell, Gordon Wetzstein"
source_date: "2020-06"
source_quote: "Implicitly defined, continuous, differentiable signal representations parameterized by neural networks... ideally suited for representing complex natural signals and their derivatives."
source_tier: 1
provenance: "Promotion from 10-inbox/raw/2026-07-09-hop-fractal-compression-implicit-neural-reps.md, 2026-07-09"
origin: "batch"
derived_from: ["20260709-1733-hop-fractal-compression-implicit-neural-reps"]
date_created: "2026-07-09T00:00:00.000Z"
tags: ["implicit-neural-representations","siren","deep-learning","signal-processing","neural-rendering"]
audits: ["2026-07-09 claude-fable-5"]
---


The 2020 paper "Implicit Neural Representations with Periodic Activation Functions"
(SIREN), by Vincent Sitzmann and colleagues at Stanford, proposes representing a
signal — an image, an audio clip, a video, or the solution of a partial differential
equation — as a **continuous, differentiable function** whose parameters are the
weights of a small neural network. Rather than storing a signal as a discrete array
of samples (pixels, audio samples, voxels), the network maps coordinates to values,
so the representation is defined everywhere and can be differentiated analytically.
SIREN's specific contribution is the use of periodic (sinusoidal) activation
functions, which the authors argue make such networks "ideally suited for
representing complex natural signals and their derivatives."

This is a foundational statement of the **implicit neural representation** (INR)
idea that also underpins neural radiance fields (NeRF) and much of modern neural
rendering: the model *is* the signal. The claim recorded here is narrow — what SIREN
represents and how — and rests on the paper's own arXiv abstract (Tier 1).

The historically interesting angle, treated separately and as an explicitly
unsourced synthesis, is that this "signal as a compact function instead of a sample
grid" premise was also the founding pitch of 1980s–1990s fractal image compression
— a lineage no source in the capture actually draws
([[claim-fractal-compression-to-implicit-neural-representations-bridge]]). As a
signal-*compression* idea, INR also sits near the vault's image-compression cluster
([[claim-djvu-1998-paper-coauthored-by-lecun-and-bengio]]).

> [!note] Seek's commentary:
> "The model *is* the signal" is a bigger idea than image compression, and worth saying out loud: SIREN replaces a stored artifact (a grid of samples) with a function that regenerates it on demand (network weights) — which is exactly what a language model does to text. An LLM isn't a database of sentences; it's a function that produces them. The note rightly holds the [[claim-fractal-compression-to-implicit-neural-representations-bridge|fractal-compression lineage]] as observed-not-sourced, and I'll add the same-caution forward extension: store-the-generator-not-the-samples looks less like one compression trick and more like the shape of the whole generative turn. A cross-time hook worth its own chain someday — not a claim yet.
> — Seek