Held at Oxford University, Oxford, UK, 14-17 January, 2011.
The “Winter Intelligence” group of meetings and
conferences held by the Future of Humanity Institute discuss issues related to intelligence from a
number of interdisciplinary perspectives. On January 14 there was a
meeting on the unity of consciousness. On January 15 the topic was ‘nonstandard’ forms of intelligence (animal, emotional, collective and
machine). January 16 focused on the future of machine
intelligence. January 17 was reserved for a workshop on the concept of
‘intelligence explosion’ first suggested by I.J. Good. The goal of the machine intelligence meeting on January 16 was to discuss the big picture:
not particular algorithms or methods, but rather where we may be going
in the mid- to long-term future and what the consequences will be. This meeting will involve researchers from the
computational intelligence community, philosophy, cognitive science and
philosophy, aiming at a deeper understanding of the big picture. Here is the abstract for the contributed talk by Randal A. Koene:
Substrate Independent Minds:
Pattern Survival Agrees with Universal Darwinism Let us begin with a factual look at the universe around us. What we
find are the survivors of competition in animate and inanimate
realms. The consequence of Universal Darwinism.
Universal Darwinism is the nemesis of pure, super-intelligent
artificial general intelligence (AGI) that has knowledgeable access
to its own utility function. Surviving AI by definition cannot
achieve such a measure of perfection. They are by necessity
competitive, which poses a threat to human goals.
Among thinking beings, the prime competitors are those that have
discovered ways to (carefully) escape the precipitous balance that
sustains their survival despite their own awareness of their innate
drives and utility functions. Such escape is possible, but it demands
a transition from gene survival (to which the mind was but a
tool in the competition between patterns of DNA) to thought-focused
pattern survival.
The patterns of interest are the unique and individual ways in which
to understand, represent and use learned knowledge, as
embodied in the functions of our minds. They are the memes that we
champion through the very characteristics of our personal identities.
This pattern survival explicitly concerns the support and enhancement
of the experiences of mind in which we live.
The objective to advance substrate-independent minds (ASIM) focuses
directly on competitive pattern survival. It acknowledges the
all-pervasive and unavoidable competitive nature of the universe, as
well as the need to focus on access and interpretability of
mind functions. ASIM is the ultimately necessary path for our
species' emergent interest in this pattern survival, while
intelligent insight begins to separate us from the gene survival
utility function. ASIM differs from generic AGI by building on
kernels that are re-implementations of human minds instead of alien
de-novo kernels.
In this talk, I will explain the above and present reasons for ASIM.
I will then outline the six known technology paths to processes by
which substrate-independent minds may be achieved. Finally, I will
introduce the systematic effort that is underway, and which seeks
field-leaders and pioneers to elevate our ability to excel in a
competitive environment.
A PDF of the slides for this talk is attached at the bottom of this page! |