Zero-Shot Privacy-Aware Text Rewriting via Iterative Tree Search

2025/09/25 by Huang, Shuo, Yuan, Xingliang, Haffari, Gholamreza +1
#Computation and Language (cs.CL) #FOS: Computer and information sciences

paper · doi:10.48550/arxiv.2509.20838

Abstract

The increasing adoption of large language models (LLMs) in cloud-based services has raised significant privacy concerns, as user inputs may inadvertently expose sensitive information. Existing text anonymization and de-identification techniques, such as rule-based redaction and scrubbing, often struggle to balance privacy preservation with text naturalness and utility. In this work, we propose a zero-shot, tree-search-based iterative sentence rewriting algorithm that systematically obfuscates or deletes private information while preserving coherence, relevance, and naturalness. Our method incrementally rewrites privacy-sensitive segments through a structured search guided by a reward model, enabling dynamic exploration of the rewriting space. Experiments on privacy-sensitive datasets show that our approach significantly outperforms existing baselines, achieving a superior balance between privacy protection and utility preservation.

Citations

• On the Vulnerability of Text Sanitization
• DP-MLM: Differentially Private Text Rewriting Using Masked Language Models
• Interpretable Preferences via Multi-Objective Reward Modeling and Mixture-of-Experts
• NAP²: A Benchmark for Naturalness and Privacy-Preserving Text Rewriting by Learning from Human
• Generating Code World Models with Large Language Models Guided by Monte Carlo Tree Search
• Unsupervised Text Style Transfer via LLMs and Attention Masking with Multi-way Interactions
• Large Language Models are Advanced Anonymizers
• Reducing Privacy Risks in Online Self-Disclosures with Language Models
• Can LLMs Keep a Secret? Testing Privacy Implications of Language Models via Contextual Integrity Theory
• Locally Differentially Private Document Generation Using Zero Shot Prompting
• GPT-4 Technical Report
• Analyzing Leakage of Personally Identifiable Information in Language Models
• Language Models are Few-Shot Learners
• Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
• ER-AE: Differentially Private Text Generation for Authorship Anonymization
• Neural Legal Judgment Prediction in English
• Personalizing Dialogue Agents: I have a dog, do you have pets too?
• A Broad-Coverage Challenge Corpus for Sentence Understanding through Inference
• Membership Inference Attacks against Machine Learning Models

Related

• NAP²: A Benchmark for Naturalness and Privacy-Preserving Text Rewriting by Learning from Human
• Locally Differentially Private Document Generation Using Zero Shot Prompting
• Text Sanitization Beyond Specific Domains: Zero-Shot Redaction & Substitution with Large Language Models
• Improving Privacy Benefits of Redaction
• DP-Rewrite: Towards Reproducibility and Transparency in Differentially Private Text Rewriting