Research map / Learning theory and algorithms / Matrix and tensor methods
Spectral Learning for Supervised Topic Models
Yong Ren, Yining Wang, Jun Zhu
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 |
|---|---|---|---|
| use joint spectral estimation instead of two-stage | spectral sLDA · Amazon movie reviews (V=5,000) | pR2: range [-0.75,0.83] -> [-2.00,1.12] | Improved |
| use spectral methods instead of Gibbs sampling | spec-slda · Synthetic data | running time (s): 1313 -> 90 | Improved |
| initialize Gibbs sampling with joint spectral result (hybrid) | Gibbs sLDA · Amazon movie reviews (alpha=0.01) | pR2 | Improved |
| use joint instead of two-stage spectral method | spec-slda · Hotel Reviews | pR2 | Improved |
| increase number of training documents | spec-slda · Synthetic data | prediction error | Improved |
| increase topic number beyond 9 | spec-slda · Amazon movie reviews | robust tensor decomposition | Hurt |
| use Gibbs sampling instead of spectral methods | sLDA · Amazon movie reviews | per-word log-likelihood | Improved |
| use more words per document | spec-slda · Synthetic data | parameter recovery | Improved |
Closest papers by meaning
- Sparse Topic Modeling via Spectral Decomposition and Thresholding (JMLR 2026 · similarity 0.89)
- A Reduction for Efficient LDA Topic Reconstruction (NeurIPS 2018 · similarity 0.87)
- E-LDA: Toward Interpretable LDA Topic Models with Strong Guarantees in Logarithmic Parallel Time (ICML 2025 · similarity 0.87)
- A Statistical Approach for Optimal Topic Model Identification (JMLR 2022 · similarity 0.87)
- Pólya Urn Latent Dirichlet Allocation: A Doubly Sparse Massively Parallel Sampler (TPAMI 2018 · similarity 0.86)
- Provable Algorithms for Inference in Topic Models (ICML 2016 · 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.
