Measurement-Consistent Langevin Corrector: A Remedy for Latent Diffusion Inverse Solvers

2026-01-08

arXiv.org

AI Summary

This work addresses the instability and artifact generation—such as blob-like distortions—in existing zero-shot inverse solvers based on Latent Diffusion Models (LDMs), which stem from dynamical inconsistency during inversion. To mitigate this, we propose the Measurement-Consistent Langevin Correction (MCLC) module, a plug-and-play correction mechanism that enforces dynamical consistency without relying on restrictive linear manifold assumptions. By directly aligning the reverse diffusion trajectory with measurement constraints through a general-purpose Langevin-based update, MCLC seamlessly integrates into mainstream LDM solvers. Extensive experiments across diverse image restoration tasks demonstrate that our approach significantly enhances both reconstruction quality and numerical stability, effectively suppressing common artifacts while maintaining compatibility with existing frameworks.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

With recent advances in generative models, diffusion models have emerged as powerful priors for solving inverse problems in each domain. Since Latent Diffusion Models (LDMs) provide generic priors, several studies have explored their potential as domain-agnostic zero-shot inverse solvers. Despite these efforts, existing latent diffusion inverse solvers suffer from their instability, exhibiting undesirable artifacts and degraded quality. In this work, we first identify the instability as a discrepancy between the solver's and true reverse diffusion dynamics, and show that reducing this gap stabilizes the solver. Building on this, we introduce Measurement-Consistent Langevin Corrector (MCLC), a theoretically grounded plug-and-play correction module that remedies the LDM-based inverse solvers through measurement-consistent Langevin updates. Compared to prior approaches that rely on linear manifold assumptions, which often do not hold in latent space, MCLC operates without this assumption, leading to more stable and reliable behavior. We experimentally demonstrate the effectiveness of MCLC and its compatibility with existing solvers across diverse image restoration tasks. Additionally, we analyze blob artifacts and offer insights into their underlying causes. We highlight that MCLC is a key step toward more robust zero-shot inverse problem solvers.

Problem

• latent diffusion models
• inverse problems
• solver instability
• artifacts
• zero-shot solvers

Innovation

• Latent Diffusion Models
• Inverse Problems
• Langevin Dynamics
• Measurement Consistency
• Zero-shot Solvers