README.md

---
license: apache-2.0
language:
- it
tags:
- legal
widget:
- text: "Modifica dell' area marina protetta denominata Cinque Terre"
---
# Gulbert-ft-ita

<!-- Provide a quick summary of what the model is/does. -->

This model can be used for multi-label classification of Italian legislative acts, according to the subject index (taxonomy) currently adopted in the Gazzetta Uffciale. The model has been obtained by fine-tuning a [BERT-XXL Italian](https://huggingface.co/dbmdz/bert-base-italian-xxl-uncased) model on a large corpus of legislative acts published in the Gazzetta Ufficiale from 1988 until early 2022. 

## Model Details

### Model Description

<!-- Provide a longer summary of what this model is. -->

- **Language(s) (NLP):** Italian
- **License:** apache-2.0
- **Finetuned from model:** https://huggingface.co/dbmdz/bert-base-italian-xxl-uncased

### Model Sources

<!-- Provide the basic links for the model. -->

- **Repository:** https://huggingface.co/dhfbk
- **Paper:** M. Rovera, A. Palmero Aprosio, F. Greco, M. Lucchese, S. Tonelli and A. Antetomaso (2023) **Italian Legislative Text Classification for Gazzetta Ufficiale**. In *Proceedings of the Fifth Natural Legal Language Workshop* (NLLP2023).
- **Demo:** https://dh-server.fbk.eu/ipzs-ui-demo/

## Uses


### Direct Use

Multi-label text classification of Italian legislative acts.


## Training Details

### Training Data

The [dataset](https://github.com/dhfbk/gazzetta-ufficiale) used for training the model can be retrieved at our [GitHub account](https://github.com/dhfbk) and is further documented in the above mentioned paper.



## Evaluation


### Results

The model achieves a micro-F1 score of 0.873, macro-F1 of 0.471 and a weighted-F1 of 0.864 on the test set (3-fold average).



## Citation

<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->

**BibTeX:**
```
@inproceedings{rovera-etal-2023-italian,
    title = "{I}talian Legislative Text Classification for Gazzetta Ufficiale",
    author = "Rovera, Marco  and
      Palmero Aprosio, Alessio  and
      Greco, Francesco  and
      Lucchese, Mariano  and
      Tonelli, Sara  and
      Antetomaso, Antonio",
    booktitle = "Proceedings of the Natural Legal Language Processing Workshop 2023",
    year = "2023",
    address = "Singapore",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2023.nllp-1.6",
    pages = "44--50"
}
```