Web demos DL4papers


Source code and data: (1) | Related publications: [under review]

This web-demo illustrates how DL4papers works with full-text manuscripts. The deep learning model was trained with the BRONCO dataset, composed of several PubMed articles on cancer. By selecting a pair of “mutation” and “drug” keywords, and after pressing the “Submit” button, all articles that have those keywords and a score higher than 0.2 are listed. Each article is ranked  in order of importance based on the strength of the relation between the terms. On click, each article is opened in a new tab, where the meaningful relations are highlighted, automatically, with shades of yellow according to the score within the full text of the paper.

Some keywords examples are:  

  • R175H – NSC319726
  • V600E – PLX4032
  • H1047R – PLX4032

Contact: Leandro Bugnon