Knowledge Graph Reasoning for Explainable Recommendation

July 1, 2020

Mentor: Ishika Singh
Project Members: Aditi Goyal, Rahul Sethi, Somya Lohani, Vansh Bansal


Knowledge Graphs connect various types of information related to items into a unified space. Different paths connecting entity pairs often carry relations of different semantics, and PGPR (Policy Guided Path Reasoning) models these with the help of high-quality user and item representations generated using the TransE graph embedding scheme.
This project -

  • highlights the importance of KGs to define and interpret the process of recommendation.
  • proposes an RL-based approach (with soft rewards, a multi-hop scoring function, and action pruning)
  • imposes a beam search algorithm to sample diverse reasoning paths and items for recommendation.
  • evaluates this method on four Amazon datasets to get explicit reasoning behind the predicted paths.

Documentation: Link