---
title: "LLM-Guided Measurement Credibility Correction for Trustworthy Industrial Process Inference"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2607.06111v1"
permalink: "https://feed.astralor.com/items/itm_2423b54fdce18ce92261304f"
published: "2026-07-07T10:21:55.000Z"
score: 79
channel: "tech"
tags: ["LLM","MCC","Trustworthy AI","Industrial Process Inference","Soft Sensing"]
---

# LLM-Guided Measurement Credibility Correction for Trustworthy Industrial Process Inference

## AI commentary

> This paper is worth noting because it cleverly uses LLMs to address a tough problem in industrial deployment—measurement credibility. For engineers working on industrial AI, predictive maintenance, or soft sensor development, the method offers a lightweight, interpretable preprocessing solution that does not rely on expensive sensor redundancy. More importantly, the idea can be extended to other scenarios relying on sensor data streams, such as smart manufacturing and energy management.

## At a glance

Industrial prediction and soft sensing often fail due to untrustworthy input measurements. Existing methods rely on numerical correlation, fault labels, or explicit process equations, which are not always available. This paper proposes LLM-Guided Measurement Credibility Correction (MCC), which converts measurement semantics from process documents into usable forms for numerical models, builds independent process references, and corrects local measurement conflicts before prediction. Across multiple complex industrial tasks, MCC achieves average relative MAE reductions of 30.7% on real-test protocols and 80.3% on controlled-corruption protocols, adding only 0.5–2.0k online parameters with inference time as low as 0.089 ms/step. This demonstrates that measurement semantics can turn process documents into lightweight pre-inference credibility correction and improve prediction accuracy.

## Key points

- Proposes LLM-Guided Measurement Credibility Correction (MCC) that leverages measurement semantics from process documents to enhance input reliability for industrial prediction.
- MCC adds only 0.5–2.0k online parameters with inference time as low as 0.089 ms/step, achieving 30.7% and 80.3% MAE reductions on real-test and controlled-corruption protocols respectively.
- Addresses the limitation of existing methods that rely on numerical correlation, fault labels, or process equations, which may not be available in complex scenarios.

## Full text

Industrial prediction and soft sensing depend on credible input measurements. In field deployment, a predictor may receive biased, delayed, stale, or derived measurements that still look plausible. Prediction can then fail before the forecasting backbone becomes the main limitation, because the input window no longer represents the real process. Sensor reconstruction, data reconciliation, and fault-tolerant soft sensing reduce this risk, but they often rely on numerical correlation, alarms, fault labels, or explicit process equations. These assumptions are not always available. A correlated variable can also be an unsafe reference when variables share instruments, derived formulas, soft-sensing chains, or control actions. The key issue is to decide before prediction which external measurements can credibly support the current measurement. To address this issue, this article proposes LLM-Guided Measurement Credibility Correction (MCC). MCC converts measurement meanings in process documents into measurement semantics usable by numerical models. It builds independent process references from semantically qualified external measurements and corrects local measurement conflicts before prediction. The predictor therefore receives a more credible input window. Across multiple complex industrial forecasting and soft-sensing tasks, +MCC achieves average relative MAE reductions of 30.7% on real-test protocols and 80.3% on controlled-corruption protocols. It adds only 0.5--2.0k online parameters, with the slowest +MCC inference time at 0.089 ms/step. These results show that measurement semantics can turn process documents into lightweight pre-inference credibility correction and improve prediction accuracy.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_2423b54fdce18ce92261304f