Automatically extracting Drug-Drug Interactions (DDIs) from text is a crucial and challenging task, particularly when multiple medications are taken concurrently. In this study, we propose a novel approach, called Enhanced Attention-driven Dynamic Graph Convolutional Network (E-ADGCN), for DDI extraction. Our model combines the Attention-driven Dynamic Graph Convolutional Network (ADGCN) with a feature fusion method and multi-task learning framework. The ADGCN effectively utilizes entity information and dependency tree information from biomedical texts to extract DDIs. The feature fusion method integrates User-Generated Content (UGC) and molecular information with drug entity information from text through dynamic routing. By leveraging external resources, our approach maximizes the auxiliary effect and improves the accuracy of DDI extraction. We evaluate the E-ADGCN model on the extended DDIExtraction2013 dataset and achieve an F1-score of 81.45%. This research contributes to the advancement of automated methods for extracting valuable drug interaction information from textual sources, facilitating improved medication management and patient safety.

Knowledge Bases (KBs) are valuable resources of human knowledge which contribute to manyapplications. However, since they are manually maintained, there is a big lag between their contents and the up-to-date information of entities. Considering a target entity in KBs, this paper investigates how Cumulative Citation Recommendation (CCR) can be used to effectively detect its worthy-citation documents in large volumes of stream data. Most global relevant models only consider semantic and temporal features of entity-document instances, which does not sufficiently exploit prior knowledge underlying entity-document instances. To tackle this problem, we present a Mixture of Experts (ME) model by introducing a latent layer to capture relationships between the entity-document instances and their latent class information. An extensive set of experiments was conducted on TREC-KBA-2013 dataset. The results show that the model can significantly achieve a better performance gain compared to state-of-the-art models in CCR.

Intuitively, not only do ratings include abundant information for learning user preferences, but also reviews ccompanied by ratings. However, most existing recommender systems take rating scores for granted and discard the wealth of information in accompanying reviews. In this paper, in order to exploit user profiles' information embedded in both ratings and reviews exhaustively, we propose a Bayesian model that links a traditional Collaborative Filtering (CF) technique with a topic model seamlessly. By employing a topic model with the review text and aligning user review topics with "user attitudes” (i.e., abstract rating patterns) over the same distribution, our method achieves greater accuracy than the traditional approach on the rating prediction task. Moreover, with review text information involved, latent user rating attitudes are interpretable and "cold-start” problem can be alleviated. This property qualifies our method for serving as a "recommender” task with very sparse datasets. Furthermore, unlike most related works, we treat each review as a document, not all reviews of each user or item together as one document, to fully exploit the reviews' information. Experimental results on 25 real-world datasets demonstrate the superiority of our model over state-of-the-art methods.