English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

Interaction between pharmacological agents can trigger unexpected adverseevents. Capturing richer and more comprehensive information about drug-druginteractions (DDI) is one of the key tasks in public health and drugdevelopment. Recently, several knowledge graph embedding approaches havereceived increasing attention in the DDI domain due to their capability ofprojecting drugs and interactions into a low-dimensional feature space forpredicting links and classifying triplets. However, existing methods only applya uniformly random mode to construct negative samples. As a consequence, thesesamples are often too simplistic to train an effective model. In this paper, wepropose a new knowledge graph embedding framework by introducing adversarialautoencoders (AAE) based on Wasserstein distances and Gumbel-Softmax relaxationfor drug-drug interactions tasks. In our framework, the autoencoder is employedto generate high-quality negative samples and the hidden vector of theautoencoder is regarded as a plausible drug candidate. Afterwards, thediscriminator learns the embeddings of drugs and interactions based on bothpositive and negative triplets. Meanwhile, in order to solve vanishing gradientproblems on the discrete representation--an inherent flaw in traditionalgenerative models--we utilize the Gumbel-Softmax relaxation and the Wassersteindistance to train the embedding model steadily. We empirically evaluate ourmethod on two tasks, link prediction and DDI classification. The experimentalresults show that our framework can attain significant improvements andnoticeably outperform competitive baselines.

Drug-Drug Interaction Prediction with Wasserstein Adversarial Autoencoder-based Knowledge Graph Embeddings