Research map / Label-efficient and robust learning / Recognition and long-tail
Don't Forget the Quantifiable Relationship between Words: Using Recurrent Neural Network for Short Text Topic Discovery
Heng-Yang Lu, Lu-Yao Xie, Ning Kang, Chong-Jun Wang, Jun-Yuan Xie
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 |
|---|---|---|---|
| add RNN-learned word-relationship and IDF priors to biterm topic model | BTM · Online Questions M=10 | topic coherence: -116.0 -> -105.3 | Improved |
| add RNN-learned word-relationship and IDF priors to biterm topic model | BTM · Online News Titles M=20 | topic coherence: -639.8 -> -602.4 | Improved |
| use biterm-based short text models instead of LDA | LDA · Online Questions M=10 | topic coherence: -216.7 -> -116.0 | Improved |
| delete topic-irrelevant biterms (d-BTM) instead of BTM | BTM · Online Questions M=10 | topic coherence: -116.0 -> -119.9 | Hurt |
| replace biterm deletion (d-BTM) with IDF probabilistic weighting | d-BTM · Online News Titles M=20 | topic coherence: -650.8 -> -602.4 | Improved |
| improve RNN perplexity used for prior learning (ablation) | RIBS-TM · Online Questions | coherence and clustering | Improved |
| add RIBS-TM priors for clustering | BTM · Online Questions and News Titles | purity and entropy | Improved |
Closest papers by meaning
- TopicRNN: A Recurrent Neural Network with Long-Range Semantic Dependency (ICLR 2017 · similarity 0.86)
- Meta-Complementing the Semantics of Short Texts in Neural Topic Models (NeurIPS 2022 · similarity 0.84)
- Recurrent Hierarchical Topic-Guided RNN for Language Generation (ICML 2020 · similarity 0.84)
- Discovering Discrete Latent Topics with Neural Variational Inference (ICML 2017 · similarity 0.84)
- Knowledge-Aware Bayesian Deep Topic Model (NeurIPS 2022 · similarity 0.84)
- textTOvec: DEEP CONTEXTUALIZED NEURAL AUTOREGRESSIVE TOPIC MODELS OF LANGUAGE WITH DISTRIBUTED COMPOSITIONAL PRIOR (ICLR 2019 · similarity 0.84)
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.
