---
format: "aidr-story-markdown/v1"
id: "e86173e11b95457c5e7a3212830abfc7971f7b6e52a2dad6e0559e0f41a85d84"
canonical_url: "https://aidr.today/e86173e1?lang=en"
title: "EDGAR: Efficient Diffusion-Driven Gaussian Avatar Refinement via Synthetic Failure"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-06T00:00:00.000Z"
category: "Research"
topics: ["inference"]
source_urls: ["https://huggingface.co/papers/2611.07909"]
summary: "EDGAR is a conditional diffusion framework for refining Gaussian avatars by correcting artifacts while preserving identity fidelity, trained on EDGAR-SET, a hybrid dataset of real and synthetic samples simulating avatar failures."
---

# EDGAR: Efficient Diffusion\-Driven Gaussian Avatar Refinement via Synthetic Failure

> [Open the canonical story](<https://aidr.today/e86173e1?lang=en>)

**Published:** 2026-11-06T00:00:00.000Z
**Category:** Research
**Topics:** inference

## Summary

EDGAR is a conditional diffusion framework for refining Gaussian avatars by correcting artifacts while preserving identity fidelity, trained on EDGAR\-SET, a hybrid dataset of real and synthetic samples simulating avatar failures\.

## Sources

- [Story source](<https://huggingface.co/papers/2611.07909>)