This is if we take the retina simulation as a model. As the present, however, not enough is known about the neocortex to allow us to simulate it in such an optimized way. But the knowledge might be available by 2004 to 2008 (as we shall see in the next section). What is required, if we are to get human-level AI with hardware power at this lower bound, is the ability to simulate 1000-neuron aggregates in a highly efficient way.
The extreme alternative, which is what we assumed in the derivation of the upper bound, is to simulate each neuron individually. The number of clock cycles that neuroscientists can expend simulating the processes of a single neuron knows of no limits, but that is because their aim is to model the detailed chemical and electrodynamic processes in the nerve cell rather than to just do the minimal amount of computation necessary to replicate those features of its response function which are relevant for the total performance of the neural net. It is not known how much of the detail that is contingent and inessential and how much needs to be preserved in order for the simulation to replicate the performance of the whole. It seems like a good bet though, at least to the author, that the nodes could be strongly simplified and replaced with simple standardized elements. It appears perfectly feasible to have an intelligent neural network with any of a large variety of neuronal output functions and time delays.
It does look plausible, however, that by the time when we know how to simulate an idealized neuron and know enough about the brain's synaptic structure that we can put the artificial neurons together in a way that functionally mirrors how it is done in the brain, then we will also be able to replace whole 1000-neuron modules with something that requires less computational power to simulate than it does to simulate all the neurons in the module individually. We might well get all the way down to a mere 1000 instructions per neuron and second, as is implied by Moravec's estimate (10^14 ops / 10^11 neurons = 1000 operations per second and neuron). But unless we can build these modules without first building a whole brain then this optimization will only be possible after we have already developed human-equivalent artificial intelligence.
If we assume the upper bound on the computational power needed to simulate the human brain, i.e. if we assume enough power to simulate each neuron individually (10^17 ops), then Moore's law says that we will have to wait until about 2015 or 2024 (for doubling times of 12 and 18 months, respectively) before supercomputers with the requisite performance are at hand. But if by then we know how to do the simulation on the level of individual neurons, we will presumably also have figured out how to make at least some optimizations, so we could probably adjust these upper bounds a bit downwards.
So far I have been talking only of processor speed, but computers need a great deal of memory too if they are to replicate the brain's performance. Throughout the history of computers, the ratio between memory and speed has remained more or less constant at about 1 byte/ops. Since a signal is transmitted along a synapse, on average, with a frequency of about 100 Hz and since its memory capacity is probably less than 100 bytes (1 byte looks like a more reasonable estimate), it seems that speed rather than memory would be the bottleneck in brain simulations on the neuronal level. (If we instead assume that we can achieve a thousand-fold leverage in our simulation speed as assumed in Moravec's estimate, then that would bring the requirement of speed down, perhaps, one order of magnitude below the memory requirement. But if we can optimize away three orders of magnitude on speed by simulating 1000-neuron aggregates, we will probably be able to cut away at least one order of magnitude of the memory requirement. Thus the difficulty of building enough memory may be significantly smaller, and is almost certainly not significantly greater, than the difficulty of building a processor that is fast enough. We can therefore focus on speed as the critical parameter on the hardware front.)
This paper does not discuss the possibility that quantum phenomena are irreducibly involved in human cognition. Hameroff and Penrose and others have suggested that coherent quantum states may exist in the microtubules, and that the brain utilizes these phenomena to perform high-level cognitive feats. The author's opinion is that this is implausible. The controversy surrounding this issue won't be entered into here; it will simply be assumed, throughout this paper, that quantum phenomena are not functionally relevant to high-level brain modelling.
In conclusion we can say that the hardware capacity for human-equivalent artificial intelligence will likely exist before the end of the first quarter of the next century, and may be reached as early as 2004. A corresponding capacity should be available to leading AI labs within ten years thereafter (or sooner if the potential of human-level AI and superintelligence is by then better appreciated by funding agencies).
Notes
Software via the bottom-up approach
... continue reading