Latent Distance and Graph Feature Models

In this exercise we study two latent distance models and some widely used graph similarity measures for uni-relational data.

Exercise Overview

• Exercise 7-1: StructuredE.
• Exercise 7-2: TransE.
• Exercise 7-3: Graph Similarity Measures.

Task

• Task 7-1: Nearest Neighbor Analysis.
• Task 7-2: Wordnet Analysis.
• Task 7-3: Combining different similarity measures.

Requirements

• Python 3.5+
• Pytorch 0.2

References

• Learning Structured Embeddings of Knowledge Bases
• Translating embeddings for modeling multi-relational data.
• The Link Prediction Problem for Social Networks