Research map / Federated learning and privacy / Privacy attacks and secure learning
Privacy-Preserving Classification of Personal Text Messages with Secure Multi-Party Computation
Devin Reich, Ariel Todoki, Rafael Dowsley, Martine De Cock, anderson nascimento
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| select 50 features by information gain instead of using all 123,482 features (ablation) | SMC logistic regression (unigrams+bigrams) · hate-speech tweets | total secure classification time (s): 5,396.8 -> 5.3 | Improved |
| feature selection to 50 features (ablation) | SMC logistic regression (unigrams+bigrams) · hate-speech tweets | accuracy: 73.8% -> 73.8% | No clear effect |
| add bigrams to unigrams as features (ablation) | SMC text classifiers · hate-speech tweets | accuracy: 72.4% -> 73.8% (LR, 50 feat.) | Improved |
| add bigrams to unigrams as features (ablation) | SMC text classifiers (500 AdaBoost stumps) · hate-speech tweets | total time (s): 13.3 -> 28.3 | Hurt |
| increase AdaBoost stumps from 50 to 500 (ablation) | SMC AdaBoost (unigrams+bigrams) · hate-speech tweets | accuracy: 73.3% -> 74.4% | Improved |
| bucketization in feature extraction | SMC feature extraction · hate-speech tweets | runtime | No clear effect |
Closest papers by meaning
- Beyond Memorization: Violating Privacy via Inference with Large Language Models (ICLR 2024 · similarity 0.83)
- Synthesizing Privacy-Preserving Text Data via Finetuning *without* Finetuning Billion-Scale LLMs (ICML 2025 · similarity 0.82)
- PrE-Text: Training Language Models on Private Federated Data in the Age of LLMs (ICML 2024 · similarity 0.82)
- Self-Refining Language Model Anonymizers via Adversarial Distillation (NeurIPS 2025 · similarity 0.82)
- Certification for Differentially Private Prediction in Gradient-Based Training (ICML 2025 · similarity 0.81)
- Private and Personalized Frequency Estimation in a Federated Setting (NeurIPS 2024 · similarity 0.81)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
