Dear colleagues,
We are delighted to announce SemEval-2026 Task 3 Track B: Dimensional Stance Analysis
Aspect-Based Sentiment Analysis (ABSA) is a widely used technique for analyzing people’s opinions and sentiments at the aspect level. However, current ABSA research predominantly adopts a coarse-grained, categorical sentiment representation (e.g., positive, negative, or neutral). This approach stands in contrast to long-established theories in psychology and affective science, where sentiment is represented along fine-grained, real-valued dimensions of valence (ranging from negative to positive) and arousal (from sluggish to excited). This valence-arousal (VA) representation has inspired the rise of dimensional sentiment analysis as an emerging research paradigm, enabling more nuanced distinctions in emotional expression and supporting a broader range of applications.
Given an utterance or post and a target entity, stance detection involves determining whether the speaker is in favor or against the target. This track reformulates stance detection as a Stance-as-DimABSA task with the following transformations:
1. The stance target is treated as an aspect.
2. Discrete stance labels are replaced with continuous VA scores.
Building on this, we introduce Dimensional Stance Analysis (DimStance), a Stance-as-DimABSA task that reformulates stance detection under the ABSA schema in the VA space. This new formulation extends ABSA beyond consumer reviews to public-issue discourse (i.e., politics and environmental protection) and also generalizes stance analysis from categorical labels to continuous VA scores. Given a text and one or more aspects (targets), predict a real-valued valence-arousal (VA) score for each aspect, reflecting the stance expressed by the speaker toward it.
———————
Languages
———————
We provide data in 5 languages, including: German (deu), English (eng), Hausa (hau), Swahili (swa), and Chinese (zho)
———————
Evaluation
———————
RMSE is used.
———————
Participation
———————
Website (checkout details):
https://github.com/DimABSA/DimABSA2026
Codabench (register and submit results)
- Track B: https://www.codabench.org/competitions/11139/
Discord (community and discussion)
https://discord.gg/xWXDWtkMzu
Google Group (official updates):
https://groups.google.com/g/dimabsa-participants
———————
Important Dates
———————
- Sample Data Ready: 15 July 2025
- Training Data Ready: 30 September 2025
- Evaluation Start: 12 January 2026
- Evaluation End: 30 January 2026
- System Description Paper Due: February 2026
- Notification to Authors: March 2026
- Camera Ready Due: April 2026
- SemEval Workshop 2026: co-located with ACL 2026 (San Diego, CA, USA)
We warmly invite the community to participate in this exciting shared task and contribute to advancing NLP research.
Best regards,
SemEval-2026 Task 3 Organizers