UPC-HUB/fuelcell-ner-re

DyGIE++ Fuel Cell NER/RE Model

Named entity recognition (NER) and relation extraction (RE) model for oxygen reduction reaction (ORR) catalyst literature in fuel cells.

Trained using DyGIE++ on manually annotated fuel cell literature using the brat annotation tool.

Paper

Information Extraction from Literature for ORR Catalyst in Fuel Cell Hein Htet, Manae Hirano, Amgad Ahmed Ali Ibrahim, Yutaka Sasaki, Ryoji Asahi Computational Materials Science, 2026

GitHub

Full pipeline (RSC scraper + this model): https://github.com/upc-hub/FuelCell-IE-Pipeline

Entity Types

Type	Description	Example
catalyst	ORR catalyst material	Fe1Co2-ZNT-900
support	Catalyst support	carbon nanotube (CNT)
additive	Additive	KOH
electrolyte	Electrolyte	Nafion
precursors	Precursor material	ZIF-8
other_material	Other materials	Pt/C
material_reference	Reference material	commercial Pt/C
property	Physical/chemical property	half-wave potential
structure	Material structure	microporous
process	Synthesis/treatment process	pyrolysis
condition	Experimental condition	900 °C
value	Numerical value with unit	0.847 V

Relation Types

Type	Description
related_to	General relationship between entities
equivalent	Material equivalence (e.g. abbreviation ↔ full name)

Usage

See the GitHub repository for predict.py which downloads this model automatically.

git clone https://github.com/upc-hub/FuelCell-IE-Pipeline.git
cd FuelCell-IE-Pipeline
conda env create -f environment.yml
conda activate dygiepp
python predict.py --input article.txt --output results/

Citation

@article{htet2026fuelcell,
  title   = {Information Extraction from Literature for ORR Catalyst in Fuel Cell},
  author  = {Hein Htet and Manae Hirano and Amgad Ahmed Ali Ibrahim
             and Yutaka Sasaki and Ryoji Asahi},
  journal = {Computational Materials Science},
  year    = {2026}
}