(Wed, Apr 13, 2022)
We started with an introduction to the Semantic Web, based on the article by Tim Berners-Lee, James Hendler, Ora Lassila from 2001, available online here.
Some relevant background can be found in the Wikipedia articles on Knowledge Representation, RDF, Google Knowledge Graph, Ontologies, OWL, …
One of the basic ideas of Web3 is decentralizing the platform-basezd internet. Instead of addressing locations, can one directly refer to data and knowledge? (Some of this technology may be blockchain based, but that need not be the case.)
An example of a knowledge base is Wikidata. To see an example of how to query this knowledge base and display the results, I used the Wiki Graph Builder which is particularly nice for transitive queries such as, for example, the students of Beethoven. To get an idea what is possible seach for a term in Wikidata (such as United States) and see what relations are available (eg shares border with[1]).
We listend to the conference presentation from 2013 A Multilingual semantic wiki based on Attempto Controlled English and Grammatical Framework. This talk develops the themes of the Semantic Web alluded to above.
The aim of the homework is to explore where the Semantic Web stands today 20 years later and whether recent advances in NLP can be used to make further progress.
Notes written while listening to Ibrahim Maalouf.
Curiously, Russia appears as a neighbour to the USA only after increasing Iterations to 4. The reason is that "shares border with" should be, but is not in the database, a symmetric relation. ↩︎