Research map / Probabilistic and scientific modeling / Molecules and proteins
Discriminative Topic Modeling with Logistic LDA
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 grouped-item topic inference instead of MLP majority voting | Logistic LDA · Twitter author annotations | accuracy: 35.0% -> 38.7% | Helped |
| Use grouped-item topic inference instead of an individual MLP | Logistic LDA · Twitter tweet annotations | accuracy: 32.4% -> 35.6% | Helped |
| Use logistic LDA instead of supervised LDA | Twitter · Twitter author annotations | accuracy: 33.1% -> 38.7% | Helped |
| Use logistic LDA instead of supervised LDA | Twitter · Twitter tweet annotations | accuracy: 25.4% -> 35.6% | Helped |
| Use grouped-item inference instead of MLP majority voting | Logistic LDA · Pinterest board labels | test accuracy: 81.3% -> 82.5% | Helped |
| Use logistic LDA instead of oh-2LSTMp | 20-Newsgroups · 20-Newsgroups | test accuracy: 86.5% -> 84.4% | Hurt |
| Use logistic LDA instead of SA-LSTM | 20-Newsgroups · 20-Newsgroups | test accuracy: 84.4% -> 84.4% | No clear effect |
| Optimize network parameters with cross-entropy rather than the loss-insensitive objective (ablation) | Logistic LDA · 20-Newsgroups | accuracy: 82.2% -> 84.4% | Helped |
Closest papers by meaning
- E-LDA: Toward Interpretable LDA Topic Models with Strong Guarantees in Logarithmic Parallel Time (ICML 2025 · similarity 0.88)
- Online Bayesian Moment Matching for Topic Modeling with Unknown Number of Topics (NeurIPS 2016 · similarity 0.87)
- A Reduction for Efficient LDA Topic Reconstruction (NeurIPS 2018 · similarity 0.87)
- Geometric Dirichlet Means Algorithm for topic inference (NeurIPS 2016 · similarity 0.85)
- Dirichlet belief networks for topic structure learning (NeurIPS 2018 · similarity 0.85)
- FASTopic: Pretrained Transformer is a Fast, Adaptive, Stable, and Transferable Topic Model (NeurIPS 2024 · similarity 0.83)
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.