Overview
The NLPerspectives workshop, now in its fifth iteration, will take place alongside the Language Resources and Evaluation Conference (LREC) in Palma de Mallorca, Spain, from May 11 to 16, 2026. This workshop aims to delve into the emerging field of perspectivist approaches in Natural Language Processing (NLP), focusing on the importance of subjective interpretations in various NLP tasks.
Background & Relevance
Traditionally, NLP has relied on the notion of a singular ‘ground truth’ label, often achieved through aggregation or statistical methods. However, as the field evolves, it increasingly encounters tasks that are subjective and contentious, such as quality estimation and abuse detection. These tasks necessitate a more nuanced understanding of multiple valid perspectives. The concept of Data Perspectivism has been proposed as a means to address this issue, leveraging human label variation to better represent user diversity and employing evaluation strategies that accommodate disagreement. This workshop will expand discussions to include methodologies for collecting non-aggregated datasets and evaluating perspectivist models in real-world applications, integrating insights from social sciences and Human-Computer Interaction (HCI).
Key Details
- Workshop Dates: May 11-16, 2026
- Location: Palau de Congressos de Palma, Palma de Mallorca, Spain
- Submission Deadline: February 27, 2026
- Website: NLPerspectives Workshop
Eligibility & Participation
The workshop welcomes submissions from researchers and practitioners across various domains. It is particularly relevant for those engaged in NLP, machine learning, and HCI, as well as interdisciplinary researchers interested in the implications of subjectivity in data.
Submission or Application Guidelines
Papers must be submitted via the START page and should include essential information about the resources utilized in the research. All submissions must adhere to ethical research policies concerning AI and include an ethics statement. The workshop accepts three types of submissions:
– Regular research papers (up to 8 pages)
– Non-archival submissions (not included in proceedings)
– Research communications (1-page abstracts of existing work)
Additionally, papers previously rejected from the ACL rolling review can be submitted if they align with the workshop’s themes and include the corresponding reviews. All submissions will undergo a double-blind peer review process.
More Information
This workshop is significant as it aligns with the growing recognition of the importance of subjectivity in NLP. By fostering discussions around participatory approaches and the evaluation of diverse perspectives, it aims to enhance the inclusivity and fairness of NLP systems. The integration of techniques from HCI and social sciences further enriches the dialogue on how to effectively implement these methodologies in practical scenarios.
Conclusion
The NLPerspectives workshop offers a unique opportunity for researchers to engage with cutting-edge discussions on subjectivity in NLP. Participants are encouraged to submit their work, share insights, and collaborate with peers in this evolving field. For more information, visit the workshop’s website or contact the organizers directly.
Category: Conferences & Workshops
Tags: nlp, perspectivism, lrec, machine learning, human-computer interaction, data collection, evaluation, fairness, multimodal ai, participatory design, social good, natural language generation
