Research map / Learning theory and algorithms / Matrix and tensor methods
Probabilistic Low-Rank Matrix Completion from Quantized Measurements
Sonia A. Bhaskar
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 bin-boundary estimation (Alg. 3) to the projected-gradient method (ablation) | Alg. 1 logistic · MovieLens 1M | RMSE: 0.8698 -> 0.8568 | Improved |
| add bin-boundary estimation (Alg. 3) to the projected-gradient method (ablation) | Alg. 1 logistic · MovieLens 1M | NMAE: 0.1590 -> 0.1559 | Improved |
| use probit instead of logistic link (ablation) | Alg. 3 · MovieLens 1M | RMSE: 0.8568 -> 0.8580 | No clear effect |
| use the quantized-observation model instead of OptSpace | Alg. 3 logistic · MovieLens 1M | RMSE: 0.8947 -> 0.8568 | Improved |
| use the quantized-observation model instead of SVT | Alg. 3 logistic · MovieLens 1M | RMSE: 0.9023 -> 0.8568 | Improved |
| use the proposed method instead of OrdRec | Alg. 3 logistic · MovieLens 1M | NMAE: 0.1705 -> 0.1559 | Improved |
| increase the number of quantization levels K (ablation) | Alg. 2 log-barrier · synthetic, n=200, p=0.2 | relative MSE | Improved |
| increase fraction of revealed entries p (ablation) | Alg. 2 log-barrier · synthetic, n=200 | relative MSE | Improved |
| use log-barrier Alg. 2 instead of trace-norm / Cai and Zhou | Alg. 2 · synthetic 1-bit, n=200, r=10, p<0.4 | relative MSE | Improved |
| use log-barrier Alg. 2 instead of trace-norm at high p, low rank | Alg. 2 · synthetic 1-bit, r=3, p>0.1 | relative MSE | No clear effect |
| estimate bin boundaries instead of knowing them (ablation) | Alg. 3 · synthetic 5-level, r=5, p>=0.2 | relative MSE | No clear effect |
| treat quantized data as real-valued (OptSpace) instead of modeling quantization | OptSpace · synthetic 5-level | relative MSE | Hurt |
Closest papers by meaning
- Robust Matrix Sensing in the Semi-Random Model (NeurIPS 2023 · similarity 0.89)
- Lower Bounds on Adaptive Sensing for Matrix Recovery (NeurIPS 2023 · similarity 0.88)
- Guarantees of a Preconditioned Subgradient Algorithm for Overparameterized Asymmetric Low-rank Matrix Recovery (ICML 2025 · similarity 0.88)
- Symmetric Matrix Completion with ReLU Sampling (ICML 2024 · similarity 0.87)
- Fast exact recovery of noisy matrix from few entries: the infinity norm approach (NeurIPS 2025 · similarity 0.87)
- A Pairwise Pseudo-likelihood Approach for Matrix Completion with Informative Missingness (NeurIPS 2024 · similarity 0.87)
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.
