UKP_Psycontrol at SemEval-2026 Task 2: Modeling Valence and Arousal Dynamics from Text

2026-04-23

AI Summary

This study addresses the modeling of emotional states and their short-term dynamics in user-generated text by proposing a unified framework that integrates large language model prompting, an Ising-inspired pairwise maximum entropy transition structure, and lightweight neural regression. The approach introduces trainable user embeddings and temporal affective trajectories to jointly predict the evolution of valence and arousal. Its key innovation lies in coupling a structured emotion transition mechanism with personalized user representations, revealing that affective dynamics are primarily driven by numerical trajectory patterns rather than textual semantics. The proposed system achieved top performance in both Subtask 1 and Subtask 2A of SemEval-2026 Task 2.

Technology Category

• Cognitive Modeling & Cognitive Systems: Affective Computing
• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.
• Humans and AI: Emotional Intelligence

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

This paper presents our system developed for SemEval-2026 Task 2. The task requires modeling both current affect and short-term affective change in chronologically ordered user-generated texts. We explore three complementary approaches: (1) LLM prompting under user-aware and user-agnostic settings, (2) a pairwise Maximum Entropy (MaxEnt) model with Ising-style interactions for structured transition modeling, and (3) a lightweight neural regression model incorporating recent affective trajectories and trainable user embeddings. Our findings indicate that LLMs effectively capture static affective signals from text, whereas short-term affective variation in this dataset is more strongly explained by recent numeric state trajectories than by textual semantics. Our system ranked first among participating teams in both Subtask 1 and Subtask 2A based on the official evaluation metric.

Problem

• valence
• arousal
• affective dynamics
• text modeling
• user-generated text

Innovation

• LLM prompting
• Maximum Entropy model
• Ising-style interactions
• affective trajectories
• trainable user embeddings