Retrieval-augmented recommender system for data anonymization methods
Abstract
The total volume of data created and consumed worldwide is growing rapidly, supporting numerous digital services that depend on detailed individual information. This growth raises privacy concerns, making anonymisation essential to address them. When choosing a technique, data practitioners need to consider the type of data, privacy regulations, and data utility requirements. The research landscape of anonymisation techniques is broad and continues to evolve. This creates a need for a system that can accept the requirements of a data-collecting use case, understand its constraints, search the scientific literature, and recommend appropriate anonymisation methods. This work proposes a recommender system for data anonymisation methods, built using Retrieval-Augmented Generation (RAG). A knowledge base of research papers on anonymisation is created and stored in a vector database. A simple RAG pipeline is set up as the starting point. Several types of improvements are tested on top of this, focusing on retrieval settings, re-ranking, query enrichment, and planning. To assess the system, a multidimensional evaluation framework is designed to cover retrieval quality, generation quality, and recommendation quality. Given the absence of a standardised benchmark for this task, a custom evaluation framework was developed from the knowledge base to allow testing. The results show that combining retrieval and generation enhancements substantially outperforms the naive baseline. The…
Read the paper
Related studies
- Knowledge Management Perspective of Generative Artificial Intelligence
- Navigating Generative Artificial Intelligence Promises and Perils for Knowledge and Creative Work
- Characterizing the evolving research on enterprise content management
- Large language models and multimodal foundation models for precision oncology
- From Generative AI to Extended Reality: Multidisciplinary Perspectives on the Challenges, Opportunities, and Future of Educational Computing
- Generative Artificial Intelligence: A New Engine for Advancing Environmental Science and Engineering
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.