SARA: An LLM-Based Approach for Systematic Literature Review with a Use Case on Classical–Quantum Integration for Secure Communication

Abstract

This work presents SARA (Semi-Automated Review and Analysis System), an approach that integrates Large Language Models (LLMs), topic modeling, and semantic search to support Systematic Literature Reviews (SLRs). As a use case, SARA was applied to the domain of integration between classical networks and secure quantum communications, covering Quantum Key Distribution (QKD) and Post-Quantum Cryptography (PQC). The system performed deduplication, pre-selection, categorization, and ranking of articles, prioritizing publications with higher potential for scientific contribution. The results indicate a predominance of studies on QKD, followed by PQC and hybrid approaches, with emphasis on telecommunications, IoT, and critical systems. SARA proved effective in identifying publications aligned with the researcher’s queries as well as conceptually related studies, demonstrating its potential to optimize SLR processes in emerging technological areas.