Research map / Federated learning and privacy / Privacy attacks and secure learning
Homomorphic Matrix Completion
Xiao-Yang Liu, Zechu Li, Xiaodong Wang
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 homomorphic encryption versions (HNN, HAM) instead of plain NN and AM | NN and AM matrix completion · Synthetic 1,000 x 1,000 low-rank | success rate (RSE <= 10^-5) | Mixed |
| increase percentage of observed entries | NN, AM and homomorphic versions · Synthetic 1,000 x 1,000 low-rank | success rate | Improved |
| increase matrix rank r | NN, AM and homomorphic versions · Synthetic 1,000 x 1,000 low-rank | success rate | Hurt |
| hide plaintext in larger space by adding k = 10 rank (ablation) | HNN and HAM · Synthetic 1,000 x 1,000 low-rank | success rate: rank r -> r + 10 | Hurt |
| use homomorphic NN/AM instead of private Frank-Wolfe (FW) | FW (epsilon=2 log(1/delta), delta=10^-6) · MovieLens10M (Top 400) | RSE | Improved |
| use homomorphic NN/AM instead of private Frank-Wolfe (FW) | FW · Netflix (Top 400) | RSE | Improved |
| use homomorphic algorithms instead of vanilla plaintext algorithms | NN and AM · MovieLens10M and Netflix | RSE | No clear effect |
| increase percentage of observed entries | NN · MovieLens10M and Netflix | RSE | Improved |
| increase percentage of observed entries | AM and FW · MovieLens10M and Netflix | RSE | Improved |
Closest papers by meaning
- Private Alternating Least Squares: Practical Private Matrix Completion with Tighter Rates (ICML 2021 · similarity 0.88)
- Online Matrix Completion: A Collaborative Approach with Hott Items (ICML 2024 · similarity 0.83)
- FicGCN: Unveiling the Homomorphic Encryption Efficiency from Irregular Graph Convolutional Networks (ICML 2025 · similarity 0.82)
- Fast exact recovery of noisy matrix from few entries: the infinity norm approach (NeurIPS 2025 · similarity 0.82)
- Optimality of Matrix Mechanism on $\ell_p^p$-metric (ICLR 2025 · similarity 0.82)
- Differentially Private Graph Diffusion with Applications in Personalized PageRanks (NeurIPS 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.
