F1: 68.2
Papers with Code · CC BY-SA 4.0
Relation Extraction
Relation Extraction is the task of predicting attributes and relations for entities in a sentence. For example, given a sentence “Barack Obama was born in Honolulu, Hawaii.”, a relation classifier aims at predicting the relation of “bornInCity”. Relation Extraction is the key component for building relation knowledge graphs, and it is of crucial significance to natural language processing applications such as structured search, sentiment analysis, question answering, and summarization.
<span class="description-source">Source: Deep Residual Learning for Weakly-Supervised Relation Extraction </span>
논문
F1: 68.2
F1: 67.4
RE Micro F1: 69.4NER Micro F1: 90.9Sentence Encoder: ALBERTCross Sentence: Yes
RE Micro F1: 63.2NER Micro F1: 88.4Sentence Encoder: ELMoCross Sentence: Yes
RE Micro F1: 62.7NER Micro F1: 87.5Sentence Encoder: ELMoCross Sentence: No
RE Micro F1: 66.49NER Micro F1: 87.61RE+ Micro F1: 62.77Sentence Encoder: BERT base
F1: 52.2
RE Micro F1: 72.7NER Micro F1: 91.3RE+ Micro F1: 70.5Sentence Encoder: T5-3B
F1: 67.6
F1: 69.4
RE Micro F1: 63.4NER Micro F1: 88.6Sentence Encoder: BERT baseCross Sentence: Yes
F1: 67.97
RE Micro F1: 55.9NER Micro F1: 82.6RE+ Micro F1: 53.6Sentence Encoder: biLSTM
RE Micro F1: 70.1
RE Micro F1: 52.1NER Micro F1: 80.8RE+ Micro F1: 49.5Sentence Encoder: -
RE Micro F1: 75.1 ± 0.49NER Micro F1: 91.6 ± 0.12RE+ Micro F1: 71.9 ± 0.56Sentence Encoder: FLAN T5 3B
F1: 69.5
F1 (1% Few-Shot): 17.0F1 (5% Few-Shot): 51.6
RE Micro F1: 66.1NER Micro F1: 88.0RE+ Micro F1: 62.4Sentence Encoder: BERT base
RE Micro F1: 73.0NER Micro F1: 91.1RE+ Micro F1: 71.1Sentence Encoder: ALBERT
F1: 65.1
F1: 67.8
F1: 67.0
RE Micro F1: 67.6NER Micro F1: 89.5RE+ Micro F1: 64.3Sentence Encoder: ALBERT