nchellwig at SemEval-2026 Task 3: Self-Consistent Structured Generation (SCSG) for Dimensional Aspect-Based Sentiment Analysis using Large Language Models

TL;DR

This paper introduces Self-Consistent Structured Generation (SCSG), a method using multiple large language model runs and majority voting to improve the reliability of dimensional aspect-based sentiment analysis across multiple languages.

Contribution

The paper proposes a novel self-consistent generation approach with efficient caching, significantly improving sentiment analysis accuracy in multilingual settings compared to single-pass methods.

Findings

Self-consistency with 15 runs outperforms single inference.

Achieved top rankings in multiple language-domain combinations.

Demonstrated statistical significance of improvements.

Abstract

We present Self-Consistent Structured Generation (SCSG) for Dimensional Aspect-Based Sentiment Analysis in SemEval-2026 Task 3 (Track A). SCSG enhances prediction reliability by executing a LoRA-adapted large language model multiple times per instance, retaining only tuples that achieve a majority consensus across runs. To mitigate the computational overhead of multiple forward passes, we leverage vLLM's PagedAttention mechanism for efficient key--value cache reuse. Evaluation across 6 languages and 8 language--domain combinations demonstrates that self-consistency with 15 executions yields statistically significant improvements over single-inference prompting, with our system (leveraging Gemma 3) ranking in the top seven across all settings, achieving second place on three out of four English subsets and first place on Tatar-Restaurant for DimASTE.