The only way of discovering the limits of the possible is to venture a little way past them into the impossible (Arthur C. Clarke's 2nd law)

Thursday 7 April 2011

The case for substrate-independent minds and whole brain emulation (extended abstract)

Randal Koene, Halcyon Molecular; carboncopies.org

Embracing competitive developments: The case for substrate-independent minds and whole brain emulation

In a real environment that is dominated by Universal Darwinism, all-pervasive competition and natural selection at every scale, the strongest requirement for the successful existence of a thinking entity is a survival-oriented self-consistent reward mechanism. Some thinking entities will learn ways to modify their reward mechanisms in accordance with new information, as a means to maximize reward over time. There is evidence that this development has already commenced in humans.

Unfortunately, there is also a causal deduction that demonstrates that a lack of a fixed, intrinsic-drive based “purpose” can lead to the adoption of a nihilistic philosophy with behavioral consequences such as catatonic passivity, self-termination or simple out-selection in competition with other entities that have reward mechanisms strongly grounded in resource competition and survival. Two important consequences of this insight are that truly general super-intelligent AI may be impossible to develop or evolve, and that human intelligence may itself exist in a niche-specific finely-tuned balance, such that most modifications could endanger survival.

Despite those constraints, when we take an end-result perspective, it is reasonable to expect that of those proportions of space-time that are dominated by intelligent entities, the largest proportions are occupied by entities that have achieved the ability to adapt their capabilities and requirements to many different environments. The make-up of the universe is affected by successfully competing patterns.

As humans, the awareness of our existence takes place within the experiences processed in our minds. We have individual interests in specific future experiences. Our individual characteristic ways of acquiring, representing and using knowledge support specific memes that affect future developments. Consequently, we have an interest in the competitive survival of the patterns of our functions of mind, more-so than we are interested in the survival of sequences of nucleotides in our DNA. Our thinking existence is the result of competitive developments aimed at gene-survival. The focus of our thinking existence, though, is not a focus merely on gene-survival. Therefore, we need to escape the precarious balance that was optimized for gene-survival in the environment of Earth's biosphere.

Seeking strategies optimized for pattern survival of mind functions is the very definition of the objective to achieve substrate-independent minds. One of the six known technology paths that may lead to the accomplishment of the objective of substrate-independent minds is whole brain emulation. Whole brain emulation is, in fact, the most conservative approach. It simply proposes that we identify the scope and resolution of those operations carried out by the brain that we experience as the functions of our minds. In the case of whole brain emulation, we then re-implement the functions in another substrate through the emulation of those operations, as implemented in the original cerebral substrate.

Whole brain emulation can solve at least one of the requirements for the development of substrate- independent minds by providing full access. Access to every operation underlying the functions of a mind enables exploration and experimentation. Experimentation will allow us to attempt gradual and tentative modifications. The aim of such a method is to discover means to modify the capabilities and even the reward functions of our minds, while sustaining survival-oriented behavior. In the following, we describe the reasoning behind the development of strategies to achieve substrate-independent minds and recent technological developments aimed at the prospect of whole brain emulation.

No comments:

Post a Comment