LexMachina 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, Data Science & Analytics, Natural Language Processing · Depth: Advanced, quick

Summary

LexMachina, a system developed for SemEval-2026 Task 2, addresses the challenge of predicting dynamic emotional valence and arousal shifts in ecological text. It specifically tackles "domain shift" and "regression to the mean," issues where encoder models struggle to differentiate stable user traits from temporal emotional changes. The system employs a DeBERTa-v3-Base backbone, utilizing a bifurcated strategy: post-hoc Isotonic Regression for valence calibration and a Domain Adversarial Neural Network (DANN) to mitigate user-bias in arousal. LexMachina achieved composite scores of r=0.645 for Valence and r=0.434 for Arousal, demonstrating the effectiveness of adversarial disentanglement in capturing nuances within longitudinal affective data.

Key takeaway

For NLP engineers developing systems to track emotional dynamics in user-generated text, LexMachina's approach offers a robust strategy. You should consider implementing a bifurcated model, combining post-hoc calibration like Isotonic Regression for valence and adversarial disentanglement via DANN for arousal, to better distinguish stable user traits from dynamic emotional shifts. This method can significantly improve generalization and accuracy in longitudinal affective data analysis.

Key insights

Adversarial disentanglement and calibrated regression effectively predict dynamic emotional states in ecological text.

Principles

• Encoder models struggle with stable traits vs. dynamic shifts.
• Adversarial disentanglement captures affective nuances.
• Post-hoc calibration improves valence prediction.

Method

LexMachina uses a DeBERTa-v3-Base backbone with Isotonic Regression for valence calibration and a Domain Adversarial Neural Network (DANN) to reduce user-bias in arousal prediction.

In practice

• Apply DANN to mitigate user-bias in time-series data.
• Use Isotonic Regression for post-hoc score calibration.
• Consider bifurcated strategies for complex predictions.

Topics

• Emotional Valence
• Emotional Arousal
• SemEval-2026 Task 2
• DeBERTa-v3-Base
• Domain Adversarial Neural Network
• Isotonic Regression
• Affective Computing

Best for: Research Scientist, AI Scientist, Machine Learning Engineer, NLP Engineer

Related on AIssential

• 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.
• The Circumplex Degeneracy Behind the Rare-Class Limit in Affect Recognition — Rare emotion recognition failures stem from affect geometry's circumplex degeneracy, not just class imbalance, demanding new distinguishing representations.
• 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.
• Exploration of symbolic methods for emotion detection for Portuguese — Symbolic emotion detection in Portuguese shows stable relative performance despite domain-specific absolute variations.
• Learning Moral Diversity: Modelling Individual Perspectives in Moral Classification of Texts — Modeling individual annotator perspectives in NLP improves moral classification and reveals subjective viewpoints.
• Do We Still Need Humans in the Loop? Comparing Human and LLM Annotation in Active Learning for Hostility Detection — LLM annotation at scale matches human aggregate performance at lower cost, but with distinct error profiles.

Open in AIssential →

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