Dual CNN for Relation Extraction with Knowledge-Based Attention and Word Embeddings

Figure 2

Textual relation representation and relative distances.