Improving Narrative Relationship Embeddings by Training with Additional Inverse-Relationship Constraints

We consider the problem of embedding character-entity relationships from the reduced semantic space of narratives, proposing and evaluating the assumption that these relationships hold under a reflection operation. We analyze this assumption and compare the approach to a baseline state-of-the-art model with a unique evaluation that simulates efficacy on a downstream clustering task with human-created labels. Although our model creates clusters that achieve Silhouette scores of -.084, outperforming the baseline -.227, our analysis reveals that the models approach the task much differently and perform well on very different examples. We conclude that our assumption might be useful for specific types of data and should be evaluated on a wider range of tasks.