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

TL;DR

This paper introduces an ensemble approach combining hybrid RoBERTa encoders and large language models for more accurate dimensional aspect-based sentiment analysis, demonstrating improved performance through ensemble learning.

Contribution

The paper presents a novel ensemble framework integrating hybrid RoBERTa and LLMs with prediction-level stacking for dimensional sentiment regression.

Findings

Ensemble learning outperforms individual models in RMSE and correlation.

Hybrid encoder improves prediction stability with combined continuous and discretized outputs.

In-context learning with LLMs enhances sentiment prediction accuracy.

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