Research map / Federated learning and privacy / Differential privacy
PetaGAIL++: Utility Optimized Private Trajectory Generation with Imitation Learning
Yingjie Ma, Bijal Bharadva, Xin Zhang, Joann Qiongna Chen
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 |
|---|---|---|---|
| Add sensitivity-aware noise allocation and WGAN-GP to PATEGAIL | PATEGAIL · Geolife noise 0.10 | DailyLoc JSD: 0.6915 -> 0.4914 | Improved |
| Add sensitivity-aware noise allocation and WGAN-GP to PATEGAIL | PATEGAIL · Geolife noise 0.10 | I-Rank JSD: 0.2698 -> 0.0607 | Improved |
| Add sensitivity-aware noise allocation and WGAN-GP to PATEGAIL | PATEGAIL · Geolife noise 0.10 | G-Rank JSD: 0.0512 -> 0.0278 | Improved |
| Add sensitivity-aware noise allocation and WGAN-GP to PATEGAIL | PATEGAIL · Geolife noise 0.01 | Radius JSD: 0.0260 -> 0.0409 | No clear effect |
| Add PATEGAIL++ to PATEGAIL under white-box membership inference | PATEGAIL · Geolife noise 0.01 | MIA AUC: 0.7208 -> 0.4962 | Improved |
| Add PATEGAIL++ to PATEGAIL under LiRA | PATEGAIL · Geolife | LiRA attack accuracy: 0.7000 -> 0.6088 | Improved |
| Add PATEGAIL++ to PATEGAIL on Telecom Shanghai | PATEGAIL · Telecom Shanghai noise 0.01 | Distance JSD: 0.1251 -> 0.0052 | Improved |
| Add PATEGAIL++ to PATEGAIL on Telecom Shanghai at large noise | PATEGAIL · Telecom Shanghai noise 1.00 | DailyLoc JSD: 0.5554 -> 0.6252 | Hurt |
| Add sensitivity-aware aggregation at user level in the LDP setting (ablation) | PATEGAIL++ (LDP) · Geolife noise 1.00 | DailyLoc JSD: 0.6084 -> 0.3758 | Improved |
| Add sensitivity-aware aggregation at user level in the LDP setting (ablation) | PATEGAIL++ (LDP) · Geolife noise 0.10 | DailyLoc JSD: 0.6003 -> 0.6916 | Hurt |
| Vary gradient penalty coefficient lambda_GP over 1, 5, 10, 15, 20 (ablation) | PATEGAIL++ · Geolife | utility metrics: 0.0216 -> 0.0512 | Mixed |
| Use only 40% of users instead of all users (ablation) | PATEGAIL++ · Geolife noise 0.01 | DailyLoc JSD: 0.6917 -> 0.6084 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- PATEGAIL — PateGail: A Privacy-Preserving Mobility Trajectory Generator with Imitation Learning (AAAI 2023) · better on 9 of 9 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Privacy without Noisy Gradients: Slicing Mechanism for Generative Model Training (NeurIPS 2024 · similarity 0.84)
- PCEvolve: Private Contrastive Evolution for Synthetic Dataset Generation via Few-Shot Private Data and Generative APIs (ICML 2025 · similarity 0.84)
- Generative Human Trajectory Recovery via Embedding-Space Conditional Diffusion (ICML 2025 · similarity 0.83)
- Privacy Amplification by Structured Subsampling for Deep Differentially Private Time Series Forecasting (ICML 2025 · similarity 0.83)
- RAPID: Retrieval Augmented Training of Differentially Private Diffusion Models (ICLR 2025 · similarity 0.83)
- Privacy-Preserving Data Release Leveraging Optimal Transport and Particle Gradient Descent (ICML 2024 · similarity 0.82)
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.
