SPAR: Session-based Pipeline for Adaptive Retrieval on Legacy File Systems

2025/12/15 by Duy Nguyen, Ha Thi Hai, Nguyen, Duy A. +5
Computer Science · Biochemistry, Genetics and Molecular Biology · Health Professions · #Information Retrieval and Search Behavior #Biomedical Text Mining and Ontologies #Electronic Health Records Systems

paper · pdf · doi:10.48550/arxiv.2512.12938

Abstract

The ability to extract value from historical data is essential for enterprise decision-making. However, much of this information remains inaccessible within large legacy file systems that lack structured organization and semantic indexing, making retrieval and analysis inefficient and error-prone. We introduce SPAR (Session-based Pipeline for Adaptive Retrieval), a conceptual framework that integrates Large Language Models (LLMs) into a Retrieval-Augmented Generation (RAG) architecture specifically designed for legacy enterprise environments. Unlike conventional RAG pipelines, which require costly construction and maintenance of full-scale vector databases that mirror the entire file system, SPAR employs a lightweight two-stage process: a semantic Metadata Index is first created, after which session-specific vector databases are dynamically generated on demand. This design reduces computational overhead while improving transparency, controllability, and relevance in retrieval. We provide a theoretical complexity analysis comparing SPAR with standard LLM-based RAG pipelines, demonstrating its computational advantages. To validate the framework, we apply SPAR to a synthesized enterprise-scale file system containing a large corpus of biomedical literature, showing improvements in both retrieval effectiveness and downstream model accuracy. Finally, we discuss design trade-offs and outline open challenges for deploying SPAR across diverse enterprise settings.

Citations

• An Agile Method for Implementing Retrieval Augmented Generation Tools in Industrial SMEs
• RankArena: A Unified Platform for Evaluating Retrieval, Reranking and RAG with Human and LLM Feedback
• Multimodal Medical Image Binding via Shared Text Embeddings
• eSapiens: A Real-World NLP Framework for Multimodal Document Understanding and Enterprise Knowledge Processing
• Improving the Accuracy and Efficiency of Legal Document Tagging with Large Language Models and Instruction Prompts
• Qwen2.5-VL Technical Report
• LLM4Tag: Automatic Tagging System for Information Retrieval via Large Language Models
• RL-based Query Rewriting with Distilled LLM for online E-Commerce Systems
• Agentic AI-Driven Technical Troubleshooting for Enterprise Systems: A Novel Weighted Retrieval-Augmented Generation Paradigm
• RAG4ITOps: A Supervised Fine-Tunable and Comprehensive RAG Framework for IT Operations and Maintenance
• A Comprehensive Survey of Retrieval-Augmented Generation (RAG): Evolution, Current Landscape and Future Directions
• RAGChecker: A Fine-grained Framework for Diagnosing Retrieval-Augmented Generation
• CBR-RAG: Case-Based Reasoning for Retrieval Augmented Generation in LLMs for Legal Question Answering
• AlphaFin: Benchmarking Financial Analysis with Retrieval-Augmented Stock-Chain Framework
• Bridging the Preference Gap between Retrievers and LLMs
• Retrieval-Augmented Generation for Large Language Models: A Survey
• LLM-TAKE: Theme Aware Keyword Extraction Using Large Language Models
• IAG: Induction-Augmented Generation Framework for Answering Reasoning Questions
• Large Language Model based Long-tail Query Rewriting in Taobao Search
• Take a Step Back: Evoking Reasoning via Abstraction in Large Language Models
• Query Rewriting for Retrieval-Augmented Large Language Models
• ImageBind: One Embedding Space To Bind Them All
• Recommender Systems with Generative Retrieval
• TagGPT: Large Language Models are Zero-shot Multimodal Taggers
• LLaMA: Open and Efficient Foundation Language Models
• Generate rather than Retrieve: Large Language Models are Strong Context Generators
• Learning Transferable Visual Models From Natural Language Supervision
• Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
• Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
• Fast Approximate Nearest Neighbor Search With The Navigating Spreading-out Graph
• Billion-scale similarity search with GPUs
• Efficient and robust approximate nearest neighbor search using Hierarchical Navigable Small World graphs
• BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

Related

• Stronger Baselines for Retrieval-Augmented Generation with Long-Context Language Models
• FunnelRAG: A Coarse-to-Fine Progressive Retrieval Paradigm for RAG
• When Retrieval Succeeds and Fails: Rethinking Retrieval-Augmented Generation for LLMs
• TC-RAG:Turing-Complete RAG's Case study on Medical LLM Systems
• Retrieval-Augmented Generation for Large Language Models: A Survey