A unified framework for interpretable elevator fault diagnosis and predictive maintenance via style-aware CoT fine-tuning

Yuhao Wang, Junjie Huang, Qiang Zhang · 2026

Read A unified framework for interpretable elevator fault diagnosis and predictive maintenance via style-aware CoT fine-tuning by Yuhao Wang, Junjie Huang, Qiang Zhang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/plos-pone-0353219.txt.

Description

Yuhao Wang, Junjie Huang, and Qiang Zhang present this 2026 PLOS One paper on elevator repair. They teach a small language model to diagnose faults and guide upkeep before breakdowns. General models struggle in this niche. They invent answers, lack trade knowledge, and cannot read raw sensor numbers. The team fine-tuned a compact model with step-by-step examples written in varied user styles. They also turned sensor readings into text the model could reason over.

The tuned model gave diagnostic answers far better than a leading general model. Its inner patterns also sorted alarm types well. It predicted vibration levels with skill too. The authors argue that for narrow trade jobs, dense domain training beats raw model size. Small tuned models, they say, offer a cheaper path to automated upkeep advice.

Catalog record

CREATORYuhao Wang, Junjie Huang, Qiang Zhang
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE55461
LICENSECC_BY
RIGHTSCC BY
SOURCEPLOS

Source: https://doi.org/10.1371/journal.pone.0353219

Browse the full library · About the Tower of Babel