New pipeline automates security and privacy taxonomy generation from app reviews

Researchers have developed TaxoScale, a novel pipeline designed to automate the construction of security and privacy taxonomies from large collections of mobile app reviews. This system addresses the scalability challenge faced by existing methods, which are typically limited to smaller scientific corpora. TaxoScale filters over 600,000 app reviews and employs Recursive Hierarchical Clustering along with LLM-based node naming to extend expert-defined taxonomies. The pipeline demonstrates superior performance compared to baseline methods across various metrics and successfully identifies new branches not present in previous taxonomies. AI

IMPACT Automates the creation of security and privacy taxonomies from large datasets, potentially improving app security analysis.

RANK_REASON The cluster describes a new research paper detailing a novel method for taxonomy generation. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• arXiv
• Hugging Face
• LLM
• Recursive Hierarchical Clustering Algorithm
• TaxoScale

• paper
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →