Using Multi-task Deep Neural Network to Explore Person Interaction from Social Media.

This work sought to identify the interactions between persons mentioned in social media to help readers construct background knowledge of a certain topic. We propose using a rich interactive tree structure to represent syntactic, contextual, and semantic information, and adopt a tree-based convolution kernel to identify segments that carry clues about personal interactions, which are then used to construct person-interaction networks. Empirical evaluations demonstrate that the proposed method is effective in detecting and extracting the interactions between persons in textual data, outperforming other existing extraction approaches. Furthermore, readers will be able to easily navigate through the network of the interactions between persons of interest that is constructed by the proposed method, and efficiently obtain insights from a massive corpus.