Research map / Probabilistic and scientific modeling / Molecules and proteins
Convex Two-Layer Modeling with Latent Structure
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 |
|---|---|---|---|
| CVX convex relaxation instead of locally trained Local model | two-layer latent structure model · English-Hebrew transliteration, 50/75/100 negatives | MRR | Helped |
| CVX instead of CRF-AE | two-layer latent structure model · MNIST inpainting k=2 | total inpainting error: 0.29 -> 0.27 | Helped |
| CVX instead of CRF-AE | two-layer latent structure model · MNIST inpainting k=3 | total inpainting error: 0.80 -> 0.79 | No clear effect |
| CVX instead of CRF-AE | two-layer latent structure model · MNIST inpainting k=4 | total inpainting error: 1.31 -> 1.28 | Helped |
| CVX instead of CRF-AE | two-layer latent structure model · MNIST inpainting p=4 | total inpainting error: 1.33 -> 1.29 | No clear effect |
| CVX instead of CRF-AE | two-layer latent structure model · MNIST inpainting p=6 | total inpainting error: 1.30 -> 1.27 | Helped |
| CVX instead of CRF-AE | two-layer latent structure model · MNIST inpainting p=8 | total inpainting error: 1.31 -> 1.28 | Helped |
| increase occluded patch size from k=2 to k=4 | CVX · MNIST inpainting | total inpainting error: 0.27 -> 1.28 | Hurt |
| increase sequence length p from 4 to 8 | CVX · MNIST inpainting | total inpainting error: 1.29 -> 1.28 | No clear effect |
| increase number of negative examples (50, 75, 100) | CVX vs Local · English-Hebrew transliteration | MRR | No clear effect |
Closest papers by meaning
- A Non-generative Framework and Convex Relaxations for Unsupervised Learning (NeurIPS 2016 · similarity 0.86)
- Convex Formulations for Training Two-Layer ReLU Neural Networks (ICLR 2025 · similarity 0.85)
- Learning latent variable structured prediction models with Gaussian perturbations (NeurIPS 2018 · similarity 0.85)
- Inductive Two-Layer Modeling with Parametric Bregman Transfer (ICML 2018 · similarity 0.85)
- Neural Networks are Convex Regularizers: Exact Polynomial-time Convex Optimization Formulations for Two-layer Networks (ICML 2020 · similarity 0.85)
- On The Concurrence of Layer-wise Preconditioning Methods and Provable Feature Learning (ICML 2025 · similarity 0.84)
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.