Research map / Learning theory and algorithms / Matrix and tensor methods
Clipped Matrix Completion: A Remedy for Ceiling Effects
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Replace squared loss with squared hinge loss on clipped entries in Fro-MC | Fro-CMC · FilmTrust Task 1 | F1: 0.27 (0.01) -> 0.35 (0.01) | Helped |
| Replace squared loss with squared hinge loss on clipped entries in Tr-MC | Tr-CMC · FilmTrust Task 1 | F1: 0.22 (0.00) -> 0.36 (0.00) | Helped |
| Replace squared loss with squared hinge loss on clipped entries in Fro-MC | Fro-CMC · FilmTrust Task 2 | F1: 0.35 (0.01) -> 0.40 (0.01) | Helped |
| Replace squared loss with squared hinge loss on clipped entries in Tr-MC | Tr-CMC · FilmTrust Task 2 | F1: 0.35 (0.01) -> 0.39 (0.00) | Helped |
| Replace squared loss with squared hinge loss on clipped entries in Fro-MC | Fro-CMC · Movielens 100K Task 1 | F1: 0.21 (0.01) -> 0.41 (0.00) | Helped |
| Replace squared loss with squared hinge loss on clipped entries in Tr-MC | Tr-CMC · Movielens 100K Task 1 | F1: 0.12 (0.00) -> 0.40 (0.00) | Helped |
| Replace squared loss with squared hinge loss on clipped entries in Fro-MC | Fro-CMC · Movielens 100K Task 2 | F1: 0.38 (0.01) -> 0.41 (0.01) | Helped |
| Replace squared loss with squared hinge loss on clipped entries in Tr-MC | Tr-CMC · Movielens 100K Task 2 | F1: 0.38 (0.00) -> 0.40 (0.00) | Helped |
| Use CMC rather than the always-positive baseline | DTr-CMC · FilmTrust Task 1 | F1: 0.41 (0.00) -> 0.47 (0.01) | Helped |
| Use CMC rather than the always-positive baseline | Fro-CMC · FilmTrust Task 2 | F1: 0.41 (0.00) -> 0.40 (0.01) | No clear effect |
| Use CMC rather than the always-positive baseline | Fro-CMC · Movielens 100K Task 1 | F1: 0.35 (0.00) -> 0.41 (0.00) | Helped |
| Use CMC rather than the always-positive baseline | Fro-CMC · Movielens 100K Task 2 | F1: 0.35 (0.00) -> 0.41 (0.01) | Helped |
Closest papers by meaning
- Dynamic matrix recovery from incomplete observations under an exact low-rank constraint (NeurIPS 2016 · similarity 0.90)
- Scalable Completion of Nonnegative Matrix with Separable Structure (AAAI 2016 · similarity 0.90)
- RGNMR: A Gauss-Newton method for robust matrix completion with theoretical guarantees (NeurIPS 2025 · similarity 0.89)
- A Pairwise Pseudo-likelihood Approach for Matrix Completion with Informative Missingness (NeurIPS 2024 · similarity 0.88)
- Fast exact recovery of noisy matrix from few entries: the infinity norm approach (NeurIPS 2025 · similarity 0.87)
- Symmetric Matrix Completion with ReLU Sampling (ICML 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.