What it does This repository is a curated bibliography of must-read papers on Neural Relation Extraction (NRE), the task of identifying semantic relationships between entities in text. It spans pattern-based methods from the mid-1990s through statistical approaches and into modern neural architectures, with direct links to each paper.

The interesting bit The maintainers—researchers at Tsinghua’s NLP lab—also built OpenNRE, so this isn’t a random list; it’s a genealogy of ideas they actually implemented. The chronological organization makes the paradigm shifts visible: watch feature engineering give way to kernels, then embeddings, then the neural architecture zoo of CNNs, RNNs, GNNs, and attention.

Key highlights

• Covers 25+ years of research, from 1995 IJCAI pattern-learning papers to 2020 EMNLP graph-neural work
• Groups papers by methodology: Pattern-Based, Statistical (Feature/Kernel/Graphical/Embedding), and Neural (Recursive, CNN, RNN, GNN, Attention)
• Includes dataset pointers: ACE 2005, TACRED, NYT, FewRel, DocRED—with most available in JSON via OpenNRE
• Three survey papers at the top for orientation
• Maintained by active researchers in the field (Tianyu Gao, Xu Han)

Caveats

• No summaries or commentary on the papers; you get titles, authors, venues, and links—nothing more
• The list appears to stop around 2020; newer transformer-era work (post-LLM) isn’t represented
• README is truncated in the source, so the full scope of later sections is unclear

Verdict Essential if you’re entering relation extraction research and need a structured map of the literature. Skip it if you want annotated explanations, code, or a survey of how large language models have reshaped the field since 2020.