Research map / Architectures and efficiency / RNNs and normalization
Dirichlet belief networks for topic structure learning
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 the simple word-distribution prior with three-layer DirBN | PFA · WS, TMN, and Twitter; especially low training-word proportions | Per-heldout-word perplexity and NPMI topic coherence | Helped |
| Replace the simple word-distribution prior with three-layer DirBN | GBN · WS, TMN, and Twitter | NPMI topic coherence | Helped |
| Use one DirBN layer instead of three (ablation) | PFA+DirBN · WS, TMN, and Twitter | NPMI topic coherence: 0.013±0.003 -> 0.021±0.005; 0.031±0.003 -> 0.046±0.003; 0.019±0.004 -> 0.024±0.001 | Hurt |
| Use one DirBN layer instead of three (ablation) | GBN+DirBN · WS, TMN, and Twitter | NPMI topic coherence: 0.015±0.005 -> 0.018±0.006; 0.032±0.002 -> 0.048±0.003; 0.021±0.004 -> 0.025±0.003 | Hurt |
| Use a three-layer hierarchy on word distributions rather than only a one-layer construction (ablation) | DirBN-equipped topic models · WS, TMN, and Twitter | Per-heldout-word perplexity | Helped |
| Use DirBN on word distributions rather than relying only on a hierarchical topic-proportion structure | Sparse-text topic models · WS, TMN, and Twitter | Per-heldout-word perplexity | Helped |
Closest papers by meaning
- Knowledge-Aware Bayesian Deep Topic Model (NeurIPS 2022 · similarity 0.90)
- Sawtooth Factorial Topic Embeddings Guided Gamma Belief Network (ICML 2021 · similarity 0.89)
- Inter and Intra Topic Structure Learning with Word Embeddings (ICML 2018 · similarity 0.89)
- TopicNet: Semantic Graph-Guided Topic Discovery (NeurIPS 2021 · similarity 0.88)
- Document Informed Neural Autoregressive Topic Models with Distributional Prior (AAAI 2019 · similarity 0.88)
- On the Affinity, Rationality, and Diversity of Hierarchical Topic Modeling (AAAI 2024 · similarity 0.87)
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.