Hi, I am actually working with Semantic Web technologies and there is a lot of going on. Linked Data has got a lot of development in the academic field. Our projects are in archaeology, cultural heritage and linguistic domains. Basically, RDF is one of the standard formats now to publish and integrate data from different places. There are also tools to generate RDF out of relational databases or any other formats (check out D2RQ, ontop, Karma, RDF extension for openRefine or rdflib python package). For triple stores, I'd recommend Blazegraph, we use it on many projects and so far for our not really big data it works fine. Other triple stores in addition to Fuseki and Virtuoso will be RDF4J, GraphDB.
The most recent hype in Semantic Web is related to word embeddings and word2vec technology, so the idea is basically to convert RDF graphs into sequences of relations and entities (RDF2vec). Shortly speaking, that would give us higher quality entity and relation predictions. Means that we will be able to extract much more semantics from large text corpora.
Hi, I am actually working with Semantic Web technologies and there is a lot of going on. Linked Data has got a lot of development in the academic field. Our projects are in archaeology, cultural heritage and linguistic domains. Basically, RDF is one of the standard formats now to publish and integrate data from different places. There are also tools to generate RDF out of relational databases or any other formats (check out D2RQ, ontop, Karma, RDF extension for openRefine or rdflib python package). For triple stores, I'd recommend Blazegraph, we use it on many projects and so far for our not really big data it works fine. Other triple stores in addition to Fuseki and Virtuoso will be RDF4J, GraphDB.
The most recent hype in Semantic Web is related to word embeddings and word2vec technology, so the idea is basically to convert RDF graphs into sequences of relations and entities (RDF2vec). Shortly speaking, that would give us higher quality entity and relation predictions. Means that we will be able to extract much more semantics from large text corpora.
Thanks for your comment!
Have you had a chance to look at Amazon Neptune?
Also, what kind of tools do you use to build ontologies?
I haven't tried Amazon Neptune. We mostly try to work with an open source software since we are in academia.
For ontologies modelling, I use protege.
Here is a list of more options.