AFourP at SemEval-2026 Task 2: Predicting Variation in Emotional Valence and Arousal over Time from Ecological Essays

2026-06-25 · Source: Paper Index on ACL Anthology · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Natural Language Processing · Depth: Expert, quick

Summary

AFourP's submission to SemEval-2026 Task 2, Subtask 1, addressed the challenge of predicting continuous emotional Valence and Arousal scores from ecological diary texts. The team's system utilized a fine-tuned RoBERTa-base model, enhanced with a single linear regression head, and processed each essay independently to capture emotional nuances. This approach yielded an rcomposite score of .679 for Valence and .466 for Arousal on the official test set. These results positioned AFourP 4th on the Subtask 1 leaderboard, highlighting a competitive performance in modeling temporal variations of emotional states from free-form textual data. The task required systems to infer emotional intensity and pleasantness from personal narratives.

Key takeaway

For NLP Engineers developing systems for continuous emotion prediction from free-form text, consider fine-tuning a RoBERTa-base model. This approach, demonstrated by AFourP's 4th place SemEval-2026 performance, shows that a simple linear regression head can effectively map textual features to Valence and Arousal scores. You should evaluate independent text processing for similar tasks to simplify model architecture and improve focus on individual narrative emotional states.

Key insights

Fine-tuning RoBERTa-base with a linear head effectively predicts emotional valence and arousal from diary texts.

Principles

• RoBERTa-base is effective for emotion prediction.
• Independent essay processing simplifies modeling.
• Linear regression heads can map embeddings to scores.

Method

The method involves fine-tuning a RoBERTa-base model with a single linear regression head. Each ecological diary essay is processed independently to predict continuous Valence and Arousal scores.

In practice

• Apply RoBERTa-base for similar text-based emotion tasks.
• Use linear heads for continuous score prediction.
• Consider independent processing for distinct text units.

Topics

• SemEval-2026
• Emotional Valence
• Emotional Arousal
• RoBERTa-base
• Natural Language Processing
• Textual Emotion Prediction
• Linear Regression

Best for: Research Scientist, AI Scientist, NLP Engineer

Related on AIssential

• Can Machines Tell Whether Gamers Love or Hate a Game? — Fine-tuning Transformer models like RoBERTa significantly improves sentiment analysis accuracy on imbalanced, complex human language datasets.
• From Affect Prediction to Affect Forecasting: Evidence for Distinct Information Sources in Longitudinal Text — Current affect prediction relies on text semantics, while future affective change forecasting benefits from numeric trajectory dynamics.
• Exploration of symbolic methods for emotion detection for Portuguese — Symbolic emotion detection in Portuguese shows stable relative performance despite domain-specific absolute variations.
• Predicting relations between SOAP note sections: The value of incorporating a clinical information model — Fine-tuning a RoBERTa-base model on manually annotated clinical data improves NER for SOAP note sections.
• Facial-Expression-Aware Prompting for Empathetic LLM Tutoring — Integrating structured facial expression data via prompt engineering significantly boosts LLM tutor empathy without retraining.
• Beyond Sentiment Classification: A Generative Framework for Emotion Intensity Evaluation in Text — Generative LLMs fine-tuned on continuous emotion intensity scores outperform classification and generalize to unseen affective dimensions.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Paper Index on ACL Anthology.