ConceptNet 5 - Real AI is around the corner

ConceptNet is a semantic neural network containing many things computers need to know about the world, especially when people understand text written by people.
The network is built of nodes representing specific layers and concepts, in the form of words or short phrases of natural language and signs of the relationship between them. These can be any things that computers need to know in order to search for information better, answer questions and understand people's goals. If you want to build your own Watson with ConceptNet, then this should be a great goal to get you started!

Documentation and API

A new release of ConceptNet (ConceptNet 5.1), described on the official wiki .
The documentation includes information about using the REST API, which allows you to:

• Retrieving data for individual nodes;
• Data requests with specified parameters;
• Requests to measure semantic distance between nodes;

ConceptNet 5 - open and free

ConceptNet 5 develops thanks to the hard work of hundreds of thousands of enthusiasts who have dedicated their time and knowledge for free. ConceptNet is free, so it is available to choose from under two Creative Commons licenses.
You can get a full-fledged ConceptNet 5 with knowledge data licensed under the Creative Commons Attribution-ShareAlike 3.0 license.
There is also a truncated version called ConceptNet 5 Basic, the license is Creative Commons Attribution 3.0. This version is free for any purpose. However, in this version there is necessarily a large amount of knowledge from Wikipedia, Wiktionary, and DBpedia, which are Attribution-ShareAlike resources.

Source codes or how to help the development of the project?

Previous versions of ConceptNet had a weak knowledge base, but there was a collection of facts and other information from people who were interested in developing the project. There is a lot more data online now than in earlier versions of ConceptNet. Information comes from various sources, so you can contribute to improving not only the state of computational knowledge, but also human knowledge in general.
To begin with, ConceptNet 5 contains almost all of the data from ConceptNet 4, also a contribution to Open Mind Common Sense.
Most of the knowledge comes from the English Wikipedia and its employees and with the help of DBpedia extracts the knowledge from the information blocks that appear on the articles. It also analyzes a large amount of content from the English Wiktionary, including synonyms and antonyms, the translation of concepts into hundreds of languages, as well as several labeled words to understand the general meaning. Vast knowledge comes from WordNet.

ConceptNet is a graph

To be precise, this is a hypergraph with nodes. Each operator in ConceptNet points to them, explaining where the information comes from and where it comes from. In the previous version of ConceptNet, it was common to use as a kind of database structures, as well as some programs for interacting with them. ConceptNet 5 is not part of software or databases; it is a network forming a hypergraph. That is, a set of vertices and edges that are represented in several formats, including JSON. You probably know better what software you want to use to interact with the network! However, you can use the original Solr index, but this is not ConceptNet, but simply a system for quickly finding knowledge and things in ConceptNet.
Some other interesting properties:

• ConceptNet graphics do not have arbitrary identifiers. Each node and claim contains all the information necessary to identify it, that is, in its URI, the network does not rely on randomly assigned IDs. The advantage of this is that if several ConceptNet branches are developed in several places, you can merge them later by simply taking the union of nodes and edges.
• ConceptNet supports data binding: you can download a list of links to the Semantic Web, through DBpedia and through RDF / OWL WordNet. For example, the concept of a cat is connected to a node on DBpedia .

ConceptNet 5 Data Download

If you want all ConceptNet data to belong to your application, you can download them ! Data in three formats:

• Flat JSON: files in which each row of a ConceptNet node is in JSON format;
• Solr JSON: a special format of JSON files that can be downloaded in the Apache Solr index for a quick search;
• CSV: simple tabular data, very convenient for command line search;

Further development of the project

Official site of the project http://conceptnet5.media.mit.edu/
ConceptNet 5 is an open source code developed under the GPLv3 license, the code is hosted on GitHub . ConceptNet is part of the Commonsense Computing Initiative , as well as collaboration between the MIT Media Lab and other laboratories, companies around the world.
The authors of this miracle: Robert Speer , on the advice of Catherine Havasi . Google Group: http://groups.google.com/group/conceptnet-users Support email: digitalintuition / soba4ka / media.mit.edu
PS And you, dear habrablitchiteli, what do you think about the trends of the project?
UPD. Thank you, newdya , for the errors seen in the post. Already fixed.