Large Language Models for Token-Efficient and Semantic-Preserving Opinion Summarization
2026-07-12 • Computation and Language
Computation and LanguageArtificial IntelligenceMachine Learning
AI summaryⓘ
The authors focus on summarizing lots of opinionated texts like reviews and social media posts. They created a method that carefully picks a small but diverse set of opinions to feed into a large language model, which then makes balanced summaries showing different viewpoints. This approach uses categories like sentiment and topics to keep the summaries accurate while using fewer words and less computing power. Their tests on product and hotel reviews, plus social media data, showed their method works better than others at covering important points fairly.
opinion summarizationlarge language modelsmultidimensional classificationsentiment analysisstratified samplingsemantic preservationtoken efficiencycontent coveragecomputational cost
Authors
Fabrizio Marozzo, Stefano Iannicelli
Abstract
Opinionated text - spanning product reviews, hotel feedback, and social posts - captures rich signals about user experiences, preferences, and concerns. However, the scale, redundancy, and imbalance of such corpora make it challenging to analyze opinions effectively, particularly when the goal is to generate summaries that remain faithful to the diversity of viewpoints expressed. This paper presents a framework that preserves semantics in LLM-based opinion summarization while minimizing token usage. We combine multidimensional classification (e.g., sentiment, topics) with a family of stratified sampling strategies to select compact yet representative subsets of opinions before prompting the LLM. Tailored prompts then produce balanced summaries that surface the salient aspects expressed in the opinions (e.g., strengths and weaknesses of products/hotels). Experiments on Amazon product reviews, Tripadvisor hotel reviews, and X/Twitter posts demonstrate that our method significantly reduces token usage and computational cost while consistently outperforming traditional AI-based and standard LLM summarization baselines in terms of content coverage, balance, and semantic preservation.