ITER: Iterative Transformer-based Entity Recognition and Relation Extraction
When extracting structured information from text, recognizing entities and extracting relationships are essential. Recent advances in both tasks generate a structured representation of the information in an autoregressive manner, a time-consuming and computationally expensive approach. This naturally raises the question of whether autoregressive methods are necessary in order to achieve comparable results. In this work, we propose ITER, an efficient encoder-based relation extraction model, that performs the task in three parallelizable steps, greatly accelerating a recent language modeling approach: ITER achieves an inference throughput of over 600 samples per second for a large model on a single consumer-grade GPU. Furthermore, we achieve state-of-the-art results on the relation extraction datasets ADE and ACE05, and demonstrate competitive performance for both named entity recognition with GENIA and CoNLL03, and for relation extraction with SciERC and CoNLL04.
PDF AbstractCode
Results from the Paper
Task | Dataset | Model | Metric Name | Metric Value | Global Rank | Benchmark |
---|---|---|---|---|---|---|
Relation Extraction | ACE 2005 | ITER | RE Micro F1 | 75.1 ± 0.49 | # 1 | |
NER Micro F1 | 91.6 ± 0.12 | # 1 | ||||
RE+ Micro F1 | 71.9 ± 0.56 | # 1 | ||||
Sentence Encoder | FLAN T5 3B | # 1 | ||||
Cross Sentence | Yes | # 1 | ||||
Relation Extraction | Adverse Drug Events (ADE) Corpus | ITER | RE+ Macro F1 | 85.6 ± 1.42 | # 1 | |
NER Macro F1 | 92.63 ± 0.89 | # 1 |