home.md

@@ -4,7 +4,7 @@
 
 ### LODEX ###
 
-Lodex - Linked Open Data Extraction Tool enables to extract large ontology vocabularies out of unstructured data such as document text or social network messages.
+Lodex - Linked Open Data Extraction Tool enables to extract large (millions of entities) ontology vocabularies out of unstructured data such as document text or social network messages.
 
 It runs as service with preconfigured freebase and DBPedia vocabularies. Nevertheless, you can create own ontologies, according to your specific project needs.
 
@@ -12,8 +12,8 @@ Lodex is not a simple entity extractor. Special features are:
 
 * adding arbitrary metatadata to an entity, not only recognizing entity types. (e.g. an id/URI)
 * dealing with huge vocabularies with millions of entities
-* find semantic entities according to label synonym lists, specified as part of your vocabulary
-* statistic disambiguation, using text similarity against entity description texts, also specified as part of your vocabulary
+* find semantic entities according to text embeddings such as BERT or matching label synonym lists, specified as part of your vocabulary
+* semantic and statistic disambiguation, using text embeddings or text similarity against entity description texts, also specified as part of your vocabulary
 * fuzzy match techniques such as Levenshtein, stopword resistent match, filtering high frequent terms with part of speech
 * search for specific entity types. These also have to be specified as part of your entity metadata inside your vocabulary