Research map / Architectures and efficiency / RNNs and normalization
LP-SparseMAP: Differentiable Relaxed Optimization for Sparse Structured Prediction
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 |
|---|---|---|---|
| Add BUDGET factors to the TREE factor (ablation) | LP-SparseMAP ListOps latent-tree tagger · ListOps test | F1: 69.12 -> 82.95 | Helped |
| Add BUDGET factors to the TREE factor (ablation) | LP-SparseMAP ListOps latent-tree tagger · ListOps test | Acc.: 68.74 -> 82.75 | Helped |
| Replace the coarse matching factor with row XOR and column AtMostOne factors (ablation) | LP-SparseMAP structured attention · SNLI test | accuracy: 84.16 -> 85.04 | Helped |
| Replace the coarse matching factor with row XOR and column AtMostOne factors (ablation) | LP-SparseMAP structured attention · MultiNLI test | accuracy: 70.36 -> 70.64 | Helped |
| Use bidirectional sequential structure instead of LP-matching (ablation) | LP-SparseMAP structured attention · MultiNLI test | accuracy: 70.64 -> 71.17 | Helped |
| Use bidirectional sequential structure instead of LP-matching (ablation) | LP-SparseMAP structured attention · SNLI test | accuracy: 85.04 -> 83.67 | Hurt |
| Use LP-SparseMAP loss instead of unstructured binary logistic loss | Multilabel classifier · bibtex test | F1: 42.28 -> 43.43 | Helped |
| Use LP-SparseMAP loss instead of structured hinge loss | Multilabel classifier · bookmarks test | F1: 33.26 -> 36.07 | Helped |
| Use structured hinge loss instead of unstructured binary logistic loss | Multilabel classifier · bibtex test | F1: 42.28 -> 37.70 | Hurt |
Closest papers by meaning
- SparseMAP: Differentiable Sparse Structured Inference (ICML 2018 · similarity 0.94)
- A Smoother Way to Train Structured Prediction Models (NeurIPS 2018 · similarity 0.87)
- Global Minimizers of $\ell^p$-Regularized Objectives Yield the Sparsest ReLU Neural Networks (NeurIPS 2025 · similarity 0.87)
- SequentialAttention++ for Block Sparsification: Differentiable Pruning Meets Combinatorial Optimization (NeurIPS 2024 · similarity 0.86)
- Random Sparse Lifts: Construction, Analysis and Convergence of finite sparse networks (ICLR 2024 · similarity 0.86)
- Dynamic Sparse Training with Structured Sparsity (ICLR 2024 · similarity 0.86)
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.