Research map / Graph learning / Knowledge and temporal graphs
Latent Relation Language Models
Hiroaki Hayashi, Zecong Hu, Chenyan Xiong, Graham Neubig
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 |
|---|---|---|---|
| Use LRLM instead of NKLM | LSTM language modeling · WikiFacts test | perplexity: 87.88 -> 82.89 | Improved |
| Use LRLM instead of NKLM | LSTM language modeling · WikiText-S test | perplexity: 55.98 -> 55.38 | Improved |
| Use LRLM instead of NKLM | LSTM language modeling · WikiText-F test | perplexity: 48.54 -> 45.70 | Improved |
| Use LRLM instead of NKLM | Transformer-XL language modeling · WikiFacts test | perplexity: 92.92 -> 76.46 | Improved |
| Use LRLM instead of NKLM | Transformer-XL language modeling · WikiText-S test | perplexity: 52.51 -> 44.98 | Improved |
| Use LRLM instead of Vanilla LM | Transformer-XL language modeling · WikiText-F test | perplexity: 33.01 -> 32.20 | Improved |
| Add extra entities extracted from article hyperlinks | LRLM on WikiFacts · WikiFacts dev | perplexity: 94.64 -> 54.60 | Improved |
| Add extra entities extracted from article hyperlinks | LRLM on WikiFacts · WikiFacts test | perplexity: 87.73 -> 51.34 | Improved |
| Treat OOV character spell-out probabilities as 1 | LRLM on WikiFacts · WikiFacts dev | perplexity: 54.60 -> 34.73 | Improved |
| Treat OOV character spell-out probabilities as 1 | LRLM on WikiFacts · WikiFacts test | perplexity: 51.34 -> 33.02 | Improved |
Closest papers by meaning
- Linearity of Relation Decoding in Transformer Language Models (ICLR 2024 · similarity 0.87)
- An Autoregressive Text-to-Graph Framework for Joint Entity and Relation Extraction (AAAI 2024 · similarity 0.86)
- MKGL: Mastery of a Three-Word Language (NeurIPS 2024 · similarity 0.86)
- DocKS-RAG: Optimizing Document-Level Relation Extraction through LLM-Enhanced Hybrid Prompt Tuning (ICML 2025 · similarity 0.85)
- Physics of Language Models: Part 3.1, Knowledge Storage and Extraction (ICML 2024 · similarity 0.85)
- A Systematic Exploration of Knowledge Graph Alignment with Large Language Models in Retrieval Augmented Generation (AAAI 2025 · 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.
