Research map / Probabilistic and scientific modeling / Molecules and proteins
Bayesian Progressive Deep Topic Model with Knowledge Informed Textual Data Coarsening Process
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 |
|---|---|---|---|
| knowledge-informed textual coarsening (ProGBN-kg/wv vs ProGBN-x) (ablation) | ProGBN · 20NG, TMN, R8, RCV1 document modeling and clustering | perplexity / km-Purity: e.g., 20NG perplexity 653 -> 614 (wv) | Helped |
| graph-enhanced decoder (ablation) | ProGBN-wv · 20NG, RCV1, R8 perplexity | perplexity: 20NG 632 -> 614; RCV1 769 -> 735; R8 427 -> 408 | Helped |
| word-embedding graph (wv) vs knowledge graph (kg) for coarsening | ProGBN · 20NG, RCV1, R8 perplexity | perplexity: 20NG 620 -> 614; RCV1 753 -> 735; R8 411 -> 408 | Helped |
| progressive coarsening + graph decoder (ProGBN-wv) | deep topic models · 20NG, TMN, R8 document clustering (all layers) | km-Purity / km-NMI: e.g., 20NG km-Purity 57.14 vs dc-ETM 48.60 | Helped |
| progressive coarsening on the bottom layer topics | ProGBN-wv · TD / TC topic quality, RCV1 TC | topic coherence | No clear effect |
| knowledge-informed coarsening on higher-layer topic diversity | ProGBN · 20NG, R8, RCV1 topic diversity (higher layers) | topic diversity | Helped |
| concatenating hierarchical latent representations | ProGBN · document clustering all layers | km-Purity: positive impact, unlike GBN/WHAI/SawETM | Helped |
Closest papers by meaning
- Knowledge-Aware Bayesian Deep Topic Model (NeurIPS 2022 · similarity 0.90)
- Alleviating "Posterior Collapse'' in Deep Topic Models via Policy Gradient (NeurIPS 2022 · similarity 0.90)
- Bayesian Deep Embedding Topic Meta-Learner (ICML 2022 · similarity 0.89)
- Dirichlet belief networks for topic structure learning (NeurIPS 2018 · similarity 0.87)
- Topic-VQ-VAE: Leveraging Latent Codebooks for Flexible Topic-Guided Document Generation (AAAI 2024 · similarity 0.86)
- On the Affinity, Rationality, and Diversity of Hierarchical Topic Modeling (AAAI 2024 · similarity 0.85)
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.