What is IEML?
- IEML (Information Economy MetaLanguage) is an open (GPL3) and free artificial metalanguage that is simultaneously a programming language, a pivot between natural languages and a semantic coordinate system. When data are categorized in IEML, the metalanguage compute their semantic relationships and distances.
- From a “social” point of view, on line communities categorizing data in IEML generate explorable ecosystems of ideas that represent their collective intelligence.
- Github.
What problems does IEML solve?
- Decompartmentalization of tags, folksonomies, taxonomies, ontologies and languages (french and english for now).
- Semantic search, automatic computing and visualization of semantic relations and distances between data.
- Giving back to the users the information that they produce, enabling reflexive collective intelligence.
Who is IEML for?
Content curators
- knowledge management
- marketing
- curation of open data from museums and libraries, crowdsourced curation
- education, collaborative learning, connectionists MOOCs
- watch, intelligence
Self-organizing on line communities
- smart cities
- collaborative teams
- communities of practice…
Researchers
- artificial intelligence
- data analytics
- humanities and social sciences, digital humanities
What motivates people to adopt IEML?
- IEML users participate in the leading edge of digital innovation, big data analytics and collective intelligence.
- IEML can enhance other AI techniques like machine learning, deep learning, natural language processing and rule-based inference.
IEML tools
IEML v.0
IEML v.0 includes…
- A dictionary of concepts whose edition is restricted to specialists but navigation and use is open to all.
- A library of tags – called USLs (Uniform Semantic Locators) – whose edition, navigation and use is open to all.
- An API allowing access to the dictionary, the library and their functionalities (semantic computing).
Intlekt v.0
Intlekt v.0 is a collaborative data curation tool that allows
– the categorization of data in IEML,
– the semantic visualization of collections of data categorized in IEML
– the publication of these collections
The prototype (to be issued in May 2018) will be mono-user but the full blown app will be social.
Who made it?
The IEML project is designed and led by Pierre Lévy.
It has been financed by the Canada Research Chair in Collective Intelligence at the University of Ottawa (2002-2016).
At an early stage (2004-2011) Steve Newcomb and Michel Biezunski have contributed to the design and implementation (parser, dictionary). Christian Desjardins implemented a second version of the dictionary. Andrew Roczniak helped for the first mathematical formalization, implemented a second version of the parser and a third version of the dictionary (2004-2016).
The 2016 version has been implemented by Louis van Beurden, Hadrien Titeux (chief engineers), Candide Kemmler (project management, interface), Zakaria Soliman and Alice Ribaucourt.
The 2017 version (1.0) has been implemented by Louis van Beurden (chief engineer), Eric Waldman (IEML edition interface, visualization), Sylvain Aube (Drupal), Ludovic Carré and Vincent Lefoulon (collections and tags management).
Dice sculpture by Tony Cragg
So if I understand correctly these upcoming tools : the dictionary and the library will be centralized (hosted by Ottawa Uni ?) and accessible via the API ?
I had originally imagined there was going to be some way to install an engine, and the natural language conversion part would be achieved via download (or periodic synchronization) of some language-specific dictionaries – for ex. like the Wikipedia Database download / currently 12 GB compressed (pages-articles.xml.bz2 – Current revisions only, no talk or user pages).
What I mean by “engine” is a stand-alone tool (i.e. distributable as a Docker container) like Apache Solr, OrientDB, etc. that developers could install on their own servers, in order to decentralize queries.
I guess I hadn’t thought about the other way around (contribution), which is restricted to specialists – for dictionnaries, at least.
For context, here’s my (ever-postponed and now a bit outdated) side-project, currently in french : https://drive.google.com/open?id=0B78zUNe7SQFeY1lsRUNZWDItTTQ
LikeLike
Yes It is an API (see Github). The stand alone engine Will be for later…
LikeLike
[…] Pour en savoir plus… […]
LikeLike
[…] “deep meaning” a new research program in artificial intelligence, based on the Information Economy MetaLanguage (IEML). I conclude this paper by evoking possible bootstrapping scenarii for the new public […]
LikeLike