Research map / Learning theory and algorithms / Optimal transport
SWIFT: Scalable Wasserstein Factorization for Sparse Nonnegative Tensors
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 |
|---|---|---|---|
| Use Wasserstein-based tensor factorization instead of distribution-specific CP losses | Tensor CP factorization · BBC News | accuracy: relative improvement ranging from $1.69\%$ to $9.65\% | Helped |
| Use Wasserstein-based tensor factorization instead of distribution-specific CP losses | Tensor CP factorization · Sutter | PR-AUC: relative improvement ranging from $5.10\%$ to $11.31\% | Helped |
| Use SWIFT factor features rather than raw input features | Lasso Logistic Regression · BBC News | accuracy: .728 $\pm$ .013 -> .759 $\pm$ .013 | Helped |
| Use SWIFT factor features rather than raw input features | Lasso Logistic Regression · Sutter | PR-AUC: .308 $\pm$ .033 -> .364 $\pm$ .063 | Helped |
| Apply SWIFT rather than the best baseline under noisy input | Tensor factorization · BBC News noisy input | accuracy: up to $15\% | Helped |
| Apply SWIFT rather than the best baseline under noisy input | Tensor factorization · Sutter noisy input | PR-AUC: up to $17\% | Helped |
| Use SWIFT instead of direct Wasserstein tensor factorization | Wasserstein tensor factorization · Sutter tensor with varying patient count | speedup: up to $293 \times$ | Helped |
| Use SWIFT instead of direct Wasserstein tensor factorization | Wasserstein tensor factorization · Varying target rank | speedup: up to $928 \times$ | Helped |
| Use SWIFT instead of direct Wasserstein tensor factorization | Wasserstein tensor factorization · Tensor size $1500 \times 500 \times 500$ | optimal transport speedup: up to $12 \times$ | Helped |
| Use sparsity-aware and rearranged computations in SWIFT | Direct Wasserstein tensor factorization · Wasserstein tensor factorization | speedup: $16 \times$ faster OT computation; up to $921 \times$ speed up | Helped |
Closest papers by meaning
- Nonparametric Sparse Tensor Factorization with Hierarchical Gamma Processes (ICML 2022 · similarity 0.89)
- Distributed Flexible Nonlinear Tensor Factorization (NeurIPS 2016 · similarity 0.89)
- Geometry-aware training of factorized layers in tensor Tucker format (NeurIPS 2024 · similarity 0.87)
- Factor Augmented Tensor-on-Tensor Neural Networks (AAAI 2025 · similarity 0.87)
- Deep Weight Factorization: Sparse Learning Through the Lens of Artificial Symmetries (ICLR 2025 · similarity 0.86)
- Implicit Regularization for Tubal Tensor Factorizations via Gradient Descent (ICML 2025 · 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.