Research map / Federated learning and privacy / Differential privacy
Two-mode Networks: Inference with as Many Parameters as Actors and Differential Privacy
Qiuping Wang, Ting Yan, Binyan Jiang, Chenlei Leng
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 |
|---|---|---|---|
| use the denoised estimator instead of the moment-based estimator | differentially private bipartite beta-model estimation · Simulation, Table 1 | 95% CI length: 1.21 -> 1.20 | Improved |
| increase the parameter magnitude c from 0.1 to 0.3 | moment estimator · Simulation (m,n)=(50,100), eps_n=log n/n^1/6 | 95% CI length: 1.21 -> 1.87 | Hurt |
| increase the parameter magnitude c from 0.1 to 0.3 | moment estimator · Simulation (m,n)=(50,100), eps_n=log n/n^1/6 | frequency estimator did not exist (x100%): 0 -> 25.54 | Hurt |
| increase the sample size from (m,n)=(50,100) to (100,200) | moment estimator · Simulation pair (1,2), eps_n=log n/n^1/6, c=0.1 | 95% CI length: 1.21 -> 0.86 | Improved |
| increase the sample size from (m,n)=(50,100) to (100,200) | moment estimator · Simulation pair (1,2), eps_n=log n/n^1/6, c=0.3 | frequency estimator did not exist (x100%): 25.54 -> 3.14 | Improved |
| use weaker privacy noise eps_n=log n/n^1/4 instead of log n/n^1/6 | moment estimator · Simulation (m,n)=(50,100), pair (1,2), c=0.3 | frequency estimator did not exist (x100%): 25.54 -> 60.43 | Hurt |
| use eps_n=log n/n^1/4 instead of log n/n^1/6 | moment estimator · Simulation (m,n)=(100,200), pair (1,2), c=0.3 | 95% CI coverage frequency (x100%): 94.08 -> 91.93 | Hurt |
| shrink the number of actor-side nodes to m=[n^1/2] (n=200,500,1000) | moment and denoised estimators · Additional simulation | estimator existence: failed to exist with probabilities larger than 96% | Hurt |
| use m=[n^3/4] with ||theta||_inf <= 0.2 log n | moment and denoised estimators · Additional simulation | estimator failure frequency: less than 4% | No clear effect |
| apply denoising algorithm to the sparse UC Irvine forum network | denoised estimator · UC Irvine forum network (899 students, 522 topics, 7089 edges) | estimator existence: non-existent every time in 1,000 releases | Hurt |
| increase epsilon (privacy budget) in the forum data analysis | moment estimator · UC Irvine forum network | 95% CI length | Improved |
Closest papers by meaning
- Efficiently Estimating Erdos-Renyi Graphs with Node Differential Privacy (NeurIPS 2019 · similarity 0.87)
- Differentially Private Community Detection for Stochastic Block Models (ICML 2022 · similarity 0.86)
- Differentially Private Learning of Undirected Graphical Models Using Collective Graphical Models (ICML 2017 · similarity 0.86)
- Controlling The Spread of Epidemics on Networks with Differential Privacy (NeurIPS 2025 · similarity 0.85)
- Universal Exact Compression of Differentially Private Mechanisms (NeurIPS 2024 · similarity 0.85)
- Differentially Private Graph Diffusion with Applications in Personalized PageRanks (NeurIPS 2024 · similarity 0.85)
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.
