Research map / Probabilistic and scientific modeling / Variational and generative sampling
Progressive EM for Latent Tree Models and Hierarchical Topic Detection
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 |
|---|---|---|---|
| Replace full-model EM with progressive EM on small submodels | HLTA parameter estimation · Nips-1k | running time (min): 42 -> 4 | Helped |
| Replace full-model EM with progressive EM on small submodels | HLTA parameter estimation · Nips-5k | running time (min): 2,020 -> 140 | Helped |
| Replace full-model EM with progressive EM on small submodels | HLTA parameter estimation · Nips-5k | topic coherence: -9.23 -> -8.04 | Helped |
| Replace full-model EM with progressive EM on small submodels | HLTA parameter estimation · Nips-1k | topic coherence: -6.23 -> -6.25 | No clear effect |
| Replace full-model EM with progressive EM on small submodels | HLTA parameter estimation · Nips-1k | per-document loglikelihood: -391 -> -390 | No clear effect |
| Use PEM-HLTA instead of nHDP | hierarchical topic detection · Nips-1k | topic coherence: -8.08 -> -6.25 | Helped |
| Use PEM-HLTA instead of nHDP | hierarchical topic detection · Nips-5k | topic coherence: -9.55 -> -8.04 | Helped |
| Use PEM-HLTA instead of nHDP | hierarchical topic detection · Nips-10k | topic coherence: -9.86 -> -8.87 | Helped |
| Use PEM-HLTA instead of nHDP | hierarchical topic detection · News-1k | topic coherence: -14.26 -> -12.30 | Helped |
| Use PEM-HLTA instead of nHDP | hierarchical topic detection · News-5k | topic coherence: -14.51 -> -13.07 | Helped |
| Use stochastic EM in the final estimation phase | PEM-HLTA · New York Times | topic coherence: -13.35 -> -12.86 | Helped |
Closest papers by meaning
- Hierarchical Methods of Moments (NeurIPS 2017 · similarity 0.85)
- Online Bayesian Moment Matching for Topic Modeling with Unknown Number of Topics (NeurIPS 2016 · similarity 0.85)
- Knowledge-Aware Bayesian Deep Topic Model (NeurIPS 2022 · similarity 0.85)
- On the Affinity, Rationality, and Diversity of Hierarchical Topic Modeling (AAAI 2024 · similarity 0.84)
- Probabilistic Non-Negative Matrix Factorization and Its Robust Extensions for Topic Modeling (AAAI 2017 · similarity 0.84)
- TopicNet: Semantic Graph-Guided Topic Discovery (NeurIPS 2021 · 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.