Welcome to
RegNLP 2025
Regulatory Natural Language Processing Workshop

The RegNLP 2025 Workshop will take place on January 20th, 2025 in conjunction with the COLING 2025 conference in Abu Dhabi, UAE.

About

The complexity, volume, and ever-changing nature of regulatory documents present unique challenges in governance, compliance, and legal frameworks across various sectors. Addressing these challenges demands specialized approaches in natural language processing (NLP) to enable effective management and utilization of regulatory content.

Recent advancements in NLP have opened new avenues for tackling these issues, specifically tailored to the domain of regulatory documents. These advancements include sophisticated techniques for document parsing, entity recognition, and automated compliance checking, which are essential for navigating the intricate landscape of regulatory requirements.

Despite these technological strides, significant open questions and challenges remain. How can NLP models better handle the dynamic and diverse nature of regulatory texts? What methods are most effective for extracting and synthesizing information from vast and complex document repositories? How can we ensure the accuracy and reliability of automated compliance tools? Moreover, what are the best practices for adapting NLP models to the highly specialized language and context of regulatory documents?

The first workshop on Regulatory Natural Language Processing (RegNLP) aims to convene a diverse group of researchers and practitioners from NLP, legal informatics, compliance, and related fields to explore these questions. We seek to share current findings, discuss challenges, and identify promising directions for future research. Most importantly, this workshop aims to foster a collaborative community dedicated to advancing NLP applications in the regulatory domain at this critical juncture.

Topics

We welcome submissions describing original work on regulatory data, as well as data with relevance to compliance and regulation, such as:

Applications of NLP to regulatory tasks including, but not limited to:

Adapting NLP methods for regulatory data including, but not limited to:

Tasks and resources:

Demos:

Industrial Research:

Interdisciplinary position papers:

!!! We have prepared a dataset for RegNLP researchers, which can be used for various regulatory NLP research tasks.

!!! The ObliQA dataset is available on GitHub : https://github.com/RegNLP/ObliQADataset.

Submission

We welcome original (unpublished) research papers in the following categories:

Note: Appendices and acknowledgements do not count towards the page limit and should adhere to the formatting guidelines detailed below.

Style & Format Guidelines

All submissions must use the official COLING templates. Templates and detailed submission guidelines are available here.

Submission Process

To submit your paper, please follow access the submission link.

Important Dates

Shared Task

Regulatory Information Retrieval and Answer Generation

RIRAG

Regulatory Information Retrieval and Answer Generation

We are organizing the Regulatory Information Retrieval and Answer Generation (RIRAG) Shared Task as part of the RegNLP workshop.

Participants will be invited to describe their system in a paper for the RegNLP workshop proceedings.

Important Dates
  • Development: 10 September 2024 - 10 November 2024
    Participants can use all provided data for Task 1 and Task 2, with test data available only for self-evaluation of models and not for determining the final awards.
  • Testing and Submission: 10 November 2024 - 20 November 2024
    Unseen questions will be released for participants to apply their developed models. Participants must submit their results within this period.
  • Evaluation: 20 November 2024 - 25 November 2024
    The organizers will evaluate all submissions, and winners will be announced at the end of this phase.
  • System Paper Submission: 25 November 2024 - 10 December 2024
    Winning teams are expected to submit a detailed system paper outlining their methodologies and findings.

Baseline System: For reference, please see the baseline system.

Join the Task: Interested participants can join the competition on Codabench.

Invited Speakers

Annie Antón

Annie Antón

Georgia Institute of Technology

Barry West

Barry West

ADGM

Workshop Organizers

Tuba Gokhan

Tuba Gokhan

MBZUAI

Kexin Wang

Kexin Wang

UKP Lab, Techinical University of Darmstadt

Iryna Gurevych

Iryna Gurevych

UKP Lab, Techinical University of Darmstadt & MBZUAI

Ted Briscoe

Ted Briscoe

MBZUAI

PROGRAM COMMITTEE

  • Sallam Abualhaija, University of Luxembourg
  • Chetan Arora, Monash University
  • Thales Bertaglia, Maastricht University
  • Travis D. Breaux, Carnegie Mellon University
  • Silvana Castano, University of Milan
  • Luigi Di Caro, University of Turin
  • Alfio Ferrara, University of Milan
  • Tunga Gungor, Bogazici University
  • Lena Held, UKP Lab, Technical University of Darmstadt
  • Timour Igamberdiev, UKP Lab, Technical University of Darmstadt
  • Prodromos Malakasiotis, Athens University of Economics and Business
  • Luisa Mich, University of Trento
  • Paulo Quaresma, University of Evora
  • Dimitrios Tsarapatsanis, University of York
  • Peter Vickers, University of Sheffield
  • Nicola Zeni, University of Trento