Open Information Extraction
In natural language processing, open information extraction is the task of generating a structured, machine-readable representation of the information in text, usually in the form of triples or n-ary propositions (Source: Wikipedia).
Papers
Showing 1–10 of 207 papers
All datasetsCaRBWiRe57OIE2016BenchIELSOIE-wikiLSOIENYTPenn TreebankWebDocOIE-healthcareDocOIE-transportationCaRB OIE benchmark (Greek Use-case)
Benchmark Results
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | DeepEx (zero-shot) | F1 | 72.6 | — | Unverified |
| 2 | DeepStruct multi-task w/ finetune | F1 | 71.3 | — | Unverified |
| 3 | Deepstruct multi-task | F1 | 71.2 | — | Unverified |
| 4 | SpanOIE [48] | F1 | 69.4 | — | Unverified |
| 5 | SpanOIE | F1 | 68.65 | — | Unverified |
| 6 | LLaMA-2-70B w/ Selected Demo & Uncertainty | F1 | 65.8 | — | Unverified |
| 7 | GPT-3.5-Turbo w/ Selected Demo & Uncertainty | F1 | 65.1 | — | Unverified |
| 8 | RnnOIE [36] | F1 | 62 | — | Unverified |
| 9 | OpenIE4 [26] | F1 | 60 | — | Unverified |
| 10 | ClausIE [9] | F1 | 59 | — | Unverified |