Many people believe that knowledge graphs are a key element of AGI: Knowledge graph as a component of AGI.
Bibligraphy:
- www.knowledgegraph.tech/ The Knowledge graph Conference
Related:
- twitter.com/yoheinakajima/status/1759107727463518702 "smallest RAG test possible of an indirect relationship on a knowledge graph"
- www.quora.com/Do-knowledge-graphs-bases-have-a-place-in-the-pursuit-of-artificial-general-intelligence-AGI-or-can-their-features-be-better-represented-in-a-learning-based-system "Do knowledge graphs / bases have a place in the pursuit of artificial general intelligence (AGI), or can their features be better represented in a learning-based system?"
- Mentions the interesting sounding "Attempto" project:
This is one of those idealistic W3C specifications with super messy implementations all over.
Reasonable introduction: www.w3.org/TR/owl2-primer/
Example: rdf/vcard.ttl.
Implemented by:
In this tutorial, we will use the Jena SPARQL hello world as a starting point. Tested on Apache Jena 4.10.0.
Basic query on rdf/vcard.ttl RDF Turtle data to find the person with full name "John Smith":
Output:
sparql --data=rdf/vcard.ttl --query=<( printf '%s\n' 'SELECT ?x WHERE { ?x <http://www.w3.org/2001/vcard-rdf/3.0#FN> "John Smith" }')---------------------------------
| x |
=================================
| <http://somewhere/JohnSmith/> |
---------------------------------To avoid writing
Output:
http://www.w3.org/2001/vcard-rdf/3.0# a billion times as queries grow larger, we can use the PREFIX syntax:
sparql --data=rdf/vcard.ttl --query=<( printf '%s\n' '
PREFIX vc: <http://www.w3.org/2001/vcard-rdf/3.0#>
SELECT ?x
WHERE { ?x vc:FN "John Smith" }
')---------------------------------
| x |
=================================
| <http://somewhere/JohnSmith/> |
---------------------------------Bibliography:
- UniProt contains some amazing examples runnable on their servers: sparql.uniprot.org/.well-known/sparql-examples/
Bibliography:
The CLI tools don't appear to be packaged for Ubuntu 23.10? Annoying... There is a package
libapache-jena-java but it doesn't contain any binaries, only Java library files.To run the CLI tools easily we can download the prebuilt:
and we can confirm it works with:
which outputs:
sudo apt install openjdk-22-jre
wget https://dlcdn.apache.org/jena/binaries/apache-jena-4.10.0.zip
unzip apache-jena-4.10.0.zip
cd apache-jena-4.10.0
export JENA_HOME="$(pwd)"
export PATH="$PATH:$(pwd)/bin"sparql -versionApache Jena version 4.10.0If your Java is too old then then running
sparql with the prebuilts fails with:
Error: A JNI error has occurred, please check your installation and try again
Exception in thread "main" java.lang.UnsupportedClassVersionError: arq/sparql has been compiled by a more recent version of the Java Runtime (class file version 55.0), this version of the Java Runtime only recognizes class file versions up to 52.0
at java.lang.ClassLoader.defineClass1(Native Method)
at java.lang.ClassLoader.defineClass(ClassLoader.java:756)
at java.security.SecureClassLoader.defineClass(SecureClassLoader.java:142)
at java.net.URLClassLoader.defineClass(URLClassLoader.java:473)
at java.net.URLClassLoader.access$100(URLClassLoader.java:74)
at java.net.URLClassLoader$1.run(URLClassLoader.java:369)
at java.net.URLClassLoader$1.run(URLClassLoader.java:363)
at java.security.AccessController.doPrivileged(Native Method)
at java.net.URLClassLoader.findClass(URLClassLoader.java:362)
at java.lang.ClassLoader.loadClass(ClassLoader.java:418)
at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:352)
at java.lang.ClassLoader.loadClass(ClassLoader.java:351)
at sun.launcher.LauncherHelper.checkAndLoadMain(LauncherHelper.java:621)Build from source is likely something like:
TODO test it.
sudo apt install maven openjdk-22-jdk
git clone https://github.com/apache/jena --branch jena-4.10.0 --depth 1
cd jena
mvn clean installIf you make the mistake of trying to run the source tree without build:
it fails with:
as per: users.jena.apache.narkive.com/T5TaEszT/sparql-tutorial-querying-datasets-error-unrecognized-option-graph
git clone https://github.com/apache/jena --branch jena-4.10.0 --depth 1
cd jena
export JENA_HOME="$(pwd)"
export PATH="$PATH:$(pwd)/apache-jena/bin"Error: Could not find or load main class arq.sparqlThey have a tutorial at: jena.apache.org/tutorials/sparql.html
Once you've done the Apache Jena CLI tools setup we can query all users with Full Name (FN) "John Smith" directly fom the rdf/vcard.ttl Turtle RDF file with the rdf/vcard.rq SPARQL query:
and that outputs:
sparql --data=rdf/vcard.ttl --query=rdf/vcard.rq---------------------------------
| x |
=================================
| <http://somewhere/JohnSmith/> |
---------------------------------Bibliography:
Hello world: stackoverflow.com/questions/16829351/is-there-a-hello-world-example-for-sparql-with-rdflib
It appears to only extract structured data from Wikipedia, not natural language, so it is kind of basic then.
Extracts a knowledge graph from Wikipedia plaintext. TODO how.
Groups concepts by hyponymy and hypernymy and meronymy and holonymy. That actually makes a lot of sense! TODO: is there a clear separation between hyponymy and meronymy?
Does not contain intermediat scientific terms, only very common ones, e.g. no mention, of "Josephson effect", "photoelectric effect"
