QuadAI at SemEval-2026 Task 3: Ensemble Learning of Hybrid RoBERTa and LLMs for Dimensional Aspect-Based Sentiment Analysis

2026-03-08

AI Summary

This study addresses the challenge of accurately and stably predicting continuous sentiment scores in dimensional aspect-based sentiment regression. To this end, we propose an ensemble learning framework that integrates a hybrid RoBERTa encoder with large language models (LLMs). Our approach jointly optimizes regression and discrete classification heads, leveraging in-context learning and ridge regression stacking to effectively unify continuous and discrete sentiment representations for the first time. This integration enables complementary strengths between LLMs and task-specific encoders. Experimental results demonstrate that the proposed method significantly reduces RMSE and improves correlation metrics on the development set, confirming its effectiveness and novelty in fine-grained sentiment analysis.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.
• Intelligent Robots: Learning & Optimization for ROB

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

We present our system for SemEval-2026 Task 3 on dimensional aspect-based sentiment regression. Our approach combines a hybrid RoBERTa encoder, which jointly predicts sentiment using regression and discretized classification heads, with large language models (LLMs) via prediction-level ensemble learning. The hybrid encoder improves prediction stability by combining continuous and discretized sentiment representations. We further explore in-context learning with LLMs and ridge-regression stacking to combine encoder and LLM predictions. Experimental results on the development set show that ensemble learning significantly improves performance over individual models, achieving substantial reductions in RMSE and improvements in correlation scores. Our findings demonstrate the complementary strengths of encoder-based and LLM-based approaches for dimensional sentiment analysis. Our development code and resources will be shared at https://github.com/aaronlifenghan/ABSentiment

Problem

• dimensional sentiment analysis
• aspect-based sentiment analysis
• sentiment regression
• ensemble learning
• RoBERTa

Innovation

• ensemble learning
• hybrid RoBERTa
• large language models
• dimensional sentiment analysis
• ridge-regression stacking