QUPID: Quantified Understanding for Enhanced Performance, Insights, and Decisions in Korean Search Engines

TL;DR

QUPID combines two small language models with different architectures to outperform large language models in search relevance, offering higher accuracy and efficiency for real-world search engines.

Contribution

The paper introduces QUPID, a novel approach that integrates generative and embedding-based small language models to improve relevance assessment and computational efficiency.

Findings

QUPID achieves higher relevance accuracy (Cohen's Kappa 0.646) than leading LLMs (0.387).

QUPID offers 60x faster inference times compared to LLMs.

QUPID improves nDCG@5 scores by 1.9% in production search pipelines.

Abstract

Large language models (LLMs) have been widely used for relevance assessment in information retrieval. However, our study demonstrates that combining two distinct small language models (SLMs) with different architectures can outperform LLMs in this task. Our approach -- QUPID -- integrates a generative SLM with an embedding-based SLM, achieving higher relevance judgment accuracy while reducing computational costs compared to state-of-the-art LLM solutions. This computational efficiency makes QUPID highly scalable for real-world search systems processing millions of queries daily. In experiments across diverse document types, our method demonstrated consistent performance improvements (Cohen's Kappa of 0.646 versus 0.387 for leading LLMs) while offering 60x faster inference times. Furthermore, when integrated into production search pipelines, QUPID improved nDCG@5 scores by 1.9%. These findings underscore how architectural diversity in model combinations can significantly enhance both search relevance and operational efficiency in information retrieval systems.