Relation Extraction (RE) is a crucial task which aims to identify and classify relations between entities in a given text, particularly, in the case of medical data, where understanding the relations between entities is essential for knowledge extraction. In the literature, Deep Learning (DL) models like BERT have shown promising results in RE tasks. Nevertheless, most of the proposed approaches do not incorporate pretraining before conducting finetuning on a new task, which could considerably improve performance. In this paper we present a novel deep learning-based approach for extracting semantic relations from unstructured medical datasets. Our approach takes place in three main phases: (1) a pretraining phase using unsupervised relation extraction techniques, (2) a fine-tuning phase employing supervised relation extraction, and (3) an inference phase for semantic relations’ deduction. Our relation extraction method is based on combining both unsupervised RE (during the pretraining phase) and supervised RE (during the fine-tuning phase). Our three phases-based approach, includes the hybridization of different BERT variants. The evaluation was carried out on the SemEval 2010 dataset. Experimental results show improvements in the evaluation metrics, which confirms the great interest of our proposal.

The Industry 4.0 concept refers to new production patterns that include new technologies, manufacturing elements, and workforce organizations. It creates highly efficient production systems that change production processes, reduce production costs and improve product quality. Quality 4.0 is an evolution of Industry 4.0, which is a modification of traditional quality control charts. In this paper, our motivation is to improve manufacturing processes as we monitor product’s quality by improving the percentage of correctly manufactured products thereby achieving efficiency. A four-layer decision-making architecture is proposed where different models and techniques are applied and a comparative study is achieved on real industrial case study: 1) data exploration layer, 2) feature engineering layer, 3) modeling layer, in which three categories of time series forecasting algorithms are experimented: statistical model (ARIMA), machine learning models (Random forest and XGBOOST) and deep learning models (Stacked LSTM and Transformer-based model), and finally 4) interpretation layer. The transformer-based model scored the best. With the classification model’s interpretation, we deducted the recommended values to monitor the product’s quality in order to reach relatively zero defects.