Extraction of drug–drug interaction using neural embedding

Abstract

Information on changes in a drug’s effect when taken in combination with a second drug, known as drug–drug interaction (DDI), is relevant in the pharmaceutical industry. DDIs can delay, decrease, or enhance absorption of either drug and thus decrease or increase their action or cause adverse effects. Information Extraction (IE) can be of great benefit in allowing identification and extraction of relevant information on DDIs. We here propose an approach for the extraction of DDI from text using neural word embedding to train a machine learning system. Results show that our system is competitive against other systems for the task of extracting DDIs, and that significant improvements can be achieved by learning from word features and using a deep-learning approach. Our study demonstrates that machine learning techniques such as neural networks and deep learning methods can efficiently aid in IE from text. Our proposed approach is well suited to play a significant role in future research.

On the road to explainable AI in drug-drug interactions prediction: A systematic review

“…In ANN, many neurons are connected in complex interconnections to solve linear or nonlinear problems. Previous studies have successfully manipulated ANN models for DDIs prediction tasks [100] , [101] . The two layers ANN model has been used in the study of Rohani et al [77] to work on a feature set of different similarity matrices collected from five different data sources.…”

Exploring the Adaptation of Recurrent Neural Network Approaches for Extracting Drug–Drug Interactions from Biomedical Text

“…We proposed a sentence level attention mechanism to determine the relevancy of a given sentence to a DDI and used a sequence learning model adopted in [29] to model the likelihood of two drug entities participating in a drug-drug interaction. We analyze this approach with the objectives of overcoming weakness in our previous study [31]. In that study a neural embedding approach based on the LSTM is used to solving the DDI extraction task.…”

“…In Fig. 4, the symbol  is a point-wise element operation and  is an element-by-element addition [31]. In the DDI recognition model, the LSTM is a module to perform binary prediction to determine DDI entities.…”

“…The embedding results are adjusted during training to minimize the loss on the supervised task. The word embedding and LSTM learning adapted in this study are similar to those used in [31]. Here, we also use cross entropy, a commonly applied distance/loss measure.…”

“…where |V| is the size of the vocabulary and output ŷ is an estimate of y. This study adopts the same application of entropy as a loss function in training neural networks, with embedding proposed by Hou and Ceesay [31].…”