This is a sample project on learning word embeddings of Named Entities from textual content. The assumption here is that the input corpus is composed of a sequence of named entities (e.g., using their labels from the DBpedia ontology) rather than raw text.
The implementation is based on the DL4J library.
- JDK 1.8 or better
- Apache Maven v3.3.1 or better
- Deep Learning for Java (DL4J) v1.0.0-beta (https://deeplearning4j.org/)
- N-Dimensional Arrays for Java (ND4J) v1.0.0-beta (https://nd4j.org)