DualAxis AI at SemEval-2026 Task 3: Dimensional Aspect-Based Sentiment Analysis

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

Summary

DualAxis AI presented a system for SemEval-2026 Task 3, focusing on Dimensional Aspect-Based Sentiment Analysis (DABSA). This approach models sentiment using continuous valence and arousal scores, moving beyond discrete polarity labels to enable fine-grained affect representation at the aspect level. SemEval 2026 Task 3 defines this setting through three subtasks, including aspect-level regression and structured extraction of aspect–opinion pairs with continuous scoring. The team implemented transformer-based baselines for all subtasks within a unified, reproducible framework. For aspect-level regression, they fine-tuned pretrained encoders in an aspect-conditioned setup to predict valence and arousal. RoBERTa-large achieved the best development performance, demonstrating average RMSEs of 0.884 for restaurant data and 0.789 for laptop data.

Key takeaway

For NLP engineers developing sentiment analysis systems, consider adopting Dimensional Aspect-Based Sentiment Analysis (DABSA) to move beyond discrete polarity labels. Your models can achieve more nuanced affect representation by predicting continuous valence and arousal scores at the aspect level. Specifically, fine-tuning pretrained transformer encoders like RoBERTa-large in an aspect-conditioned setup offers a robust baseline, as demonstrated by RMSEs of 0.884 and 0.789 on restaurant and laptop data, respectively. This approach provides richer insights than traditional methods.

Key insights

Dimensional Aspect-Based Sentiment Analysis (DABSA) uses continuous valence and arousal for fine-grained, aspect-level affect representation.

Principles

• Continuous scores offer fine-grained sentiment.
• Transformer encoders are effective for DABSA.
• Aspect-conditioned setups improve regression.

Method

The method involves fine-tuning pretrained transformer encoders in an aspect-conditioned setup to predict continuous valence and arousal scores for aspect-level sentiment regression.

In practice

• Apply RoBERTa-large for DABSA tasks.
• Use continuous scores for nuanced sentiment.
• Implement aspect-conditioned fine-tuning.

Topics

• Dimensional Sentiment Analysis
• Aspect-Based Sentiment Analysis
• SemEval 2026
• RoBERTa-large
• Transformer Models
• Valence-Arousal Model

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

Related on AIssential

• Aspect-Based Sentiment Analysis in the Accommodation Domain Using the BERTimbau Model — BERTimbau model fine-tuning strategies for Portuguese ABSA show trade-offs between recall and precision.
• 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.
• Teaching Machines to “Feel” Tweets — Domain-native pre-trained models are crucial for accurate sentiment classification in ambiguous social media language.
• DiffuSent: Towards a Unified Diffusion Framework for Aspect-Based Sentiment Analysis — DiffuSent unifies ABSA subtasks via non-auto-regressive boundary denoising, improving accuracy and speed for multi-word sentiment extraction.
• AspectRAG: An Architecture of Retrieval and Generation for Aspect-Based Sentiment Analysis — AspectRAG leverages aspect-guided retrieval and generation for unsupervised, high-performance sentiment analysis in Portuguese.
• Data augmented bootstrap: Unifying confidence interval construction by approximate invariance — Data Augmented Bootstrap (DAB) unifies confidence interval construction by leveraging approximate data invariances, extending classical and modern methods.

Open in AIssential →

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