13 avril 2009

IEML : from Social Computing to Reflexive Collective Intelligence

IEML-Levy.pdf (Objet application/pdf)

Extraits choisis :

"In cyberspace, for the first time in human history, our species is growing a universally interconnected common memory where ubiquitous data can be accessed and transformed by automatic symbol manipulators.

my hypothesis is that the main developments into the full symbolic and cognitive exploitation of the global
digital memory are still to come.

As shown by recent R&D activity addressing the improvement of social tagging [15], or the construction of a "Web of data" organized by ontologies[19, 26], the augmentation of our collective categorization power is a key issue in the development of new symbolic systems.

My main hypothesis is that natural languages, as well as notation systems invented before the 21st century, are not appropriate for the current and future scale of the social categorization process. They are not fit to exploit the new interconnected global digital memory and its unprecedented power. Natural languages are in accord with human brain processing, they were not designed to be automatically manipulated.
Old notation and writing systems correspond to heavy and slow physical storage and retrieval processes, and not to ubiquitous high speed automatic computing.

The IEML research program suggests that we should adopt a symbolic system for the notation and manipulation of concepts designed from the very beginning for massively distributed social computation in a technically interconnected global memory

IEML is a symbolic system for the notation of meaning that is “semantic content oriented” rather than “instruction oriented” like programming languages or “format oriented” like data standards. In that sense, it is closer to natural languages than to computer languages.

But there are also several differences between IEML and natural languages.
First, it is an artificial language that is mainly designed to be written and read instead of being spoken.
Secondly, natural languages have not evolved to be processed by machines but by human brains in cultural, social and emotional contexts. Unlike them, the syntax of IEML is essentially regular and IEML semantics harnesses as much as possible this : IEML is a putative language for the interface between the human mind and the computing power of cyberspace.

As the IEML dictionary can provide *tags for diverse natural languages, IEML could also be used to establish automatic correspondances between *tags expressed in
different natural languages. Once expressed in IEML, a complex concept - the meaning of a *tag - can be automatically translated to any natural language supported by the IEML dictionary.

In addition to tranparent information exchange and semantic interoperability, the IEML research
program pursues other interdependant goals that can be divided into three main categories:
1) the establishment
of a consistent and transparent semantic addressing system,
2) the development of a powerful notation of meaning for exploiting the computing power and the immense storage capacity of cyberspace,
3) the provision of a symbolic tool for self-observation and self-reference of collective intelligence.

In my judgement, the main obstacle that prevents human collective intelligence from crossing the next cognitive threshold is the current absence of systematic self-awareness...According to this hypothesis, the highest goal of the IEML research program is to provide a symbolic framework for the making of digital tools that can help human collective intelligence observe its own activity in cyberspace, and therefore improve human development.

As forms change over time, it becomes possible to represent, analyze, synthesize, compare and
simulate the dynamics of collective intelligence in semantic space."

Aucun commentaire: