REflex: Flexible Framework for Relation Extraction in Multiple Domains

Abstract

Systematic comparison of methods for relation extraction (RE) is difficult because many experiments in the field are not described precisely enough to be completely reproducible and many papers fail to report ablation studies that would highlight the relative contributions of their various combined techniques. In this work, we build a unifying framework for RE, applying this on three highly used datasets (from the general, biomedical and clinical domains) with the ability to be extendable to new datasets. By performing a systematic exploration of modeling, pre-processing and training methodologies, we find that choices of pre-processing are a large contributor performance and that omission of such information can further hinder fair comparison. Other insights from our exploration allow us to provide recommendations for future research in this area.

accepted by BioNLP 2019 at the Association of Computation Linguistics 2019

References in corpus (10)

• BioBERT: a pre-trained biomedical language representation model for biomedical text mining
• Practical Bayesian Optimization of Machine Learning Algorithms
• Natural Language Processing (almost) from Scratch
• An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling
• Comparative Study of CNN and RNN for Natural Language Processing
• Hyperparameter Search in Machine Learning
• Classifying Relations by Ranking with Convolutional Neural Networks
• Classifying medical relations in clinical text via convolutional neural networks
• Semantic Relation Classification via Convolutional Neural Networks with Simple Negative Sampling
• A hybrid deep learning approach for medical relation extraction