Lecturer's Précis - Reader (1969)
Copyright Notice: This material was written and published in Wales by Derek J. Smith (Chartered Engineer). It forms part of a multifile e-learning resource, and subject only to acknowledging Derek J. Smith's rights under international copyright law to be identified as author may be freely downloaded and printed off in single complete copies solely for the purposes of private study and/or review. Commercial exploitation rights are reserved. The remote hyperlinks have been selected for the academic appropriacy of their contents; they were free of offensive and litigious content when selected, and will be periodically checked to have remained so. Copyright © 2010, High Tower Consultants Limited.
First published online 09:30 BST 15th August 2003, Copyright Derek J. Smith (Chartered Engineer). This version [HT.1 - transfer of copyright] dated 18:00 14th January 2010
Reader's (1969) "Steps Towards Genuine Artificial Intelligence"
Reader begins by introducing the Turing Test (TT) in its original guise as an "impersonation test", and by explaining its importance as a philosophical exercise. He reminds us that the TT equates intelligence with the ability to impersonate during a conversational exchange, but goes on to suggest that this is to ignore other equally important areas of cognition .....
"Conversation is only one aspect of human behaviour that is essentially human. Other aspects are arithmetical manipulation, the control of machinery, the playing of games (such as chess), problem solving, and, more generally, artistic creation. [.....] Whilst these impersonations of human intellectual activities do not prove that the machines are intelligent, an intelligent machine would have to be able to perform these tasks [and] must continue to sustain the impersonation." (p280)
It is the ability to sustain the impersonation over time which is important, and the only way of achieving this is to be what Reader calls "open-ended" - the machine must be able to learn and never stop learning.
With this in mind, Reader proceeds to summarise the general area of animal learning, from which one observation in particular stands out ......
"..... animals seem to recognise only some combinations (patterns) of their primary sensory information [citation] and produce only comparatively few classes of response (such as eating, fleeing, attacking, pursuing, resting, scratching, etc.)" (p286)
This implies that animals must be capable (a) of pattern recognition, (b) of storing a repertoire of responses, and (c) of learning how best to relate the latter to the former in the light of experience. He then drew a flow diagram showing how he saw these three processes interacting .....
Figure 1 - Reader's (1969) "Robotic System": Here are Reader's proposals for a basic robotic processing architecture. Note the position of the three basic processing modules, namely the Pattern Recogniser [buff, bottom left], the Response Generator [buff, bottom right], and the General Learning Machine [mauve, top] discussed above. Note also the general clockwise flow of information. And note especially the "Inhibition Reinforcing Signal", which serves to signal "pain" [see Endnotes].
Redrawn from Reader (1969, p287), but rotated 90o counter-clockwise, and with additional colour coding. This graphic Copyright © 2003, Derek J. Smith.
EVALUATION: What Reader is giving us is an inverted-U shaped control hierarchy not dissimilar to Lichtheim's (1885) "House". This is therefore not really that novel, but the addition of the pain pathway is, and deserves credit.
Reader, A.V. (1969). Steps towards genuine artificial intelligence. Acta Psychologica, 29:279-289.
This is how we summarised the role of a pleasure-pain signalling system in Smith (1996; Chapter 6).....
"Pleasure-Pain Signalling: [.....] How does [a cell assembly memory] know that a particular external occurrence is worth remembering, and how does it know whether to remember it favourably or unfavourably? What is it, in other words, which gives memory its adaptive value? [//] One of the main workers here has been the British anatomist John Young, whose research into memory phenomena in cephalopods [..... suggests] that, at least in the octopus's vertical lobe and the mammalian cerebral cortex, memory is both everywhere and nowhere in particular.' (Boycott, 1965; emphasis added.) [//] It is the words 'everywhere and nowhere' which are the most significant, because they drew the Naples team strongly towards the idea of neuronal net memory mechanisms. Engrams laid down in widely distributed networks of neurons would behave in precisely the required fashion. But such widely distributed networks would only function at all if each part of the network could somehow be kept informed as to the good-bad nature of the current input. That is to say, it was not enough merely to recognise a stimulus: engrams needed also to be coded either as to-be-approached or as to-be-avoided [..... and] the necessary coding could only take place, Young argued, if there existed results indicator pathways in the CNS capable of tagging each engram with some sort of pleasure-pain evaluation. [.....] Young (1964) cites work by Maldonado (1963) which showed how an array of many "memory units" could be controlled by a single what-he-called 'noci-hedono' receptor system. This latter system served to tell all the others whether what they were doing was a good idea or not, that is to say, whether it was nasty and to be avoided, or nice and to be repeated. And much the same idea has resurfaced recently in the work of Gerald Edelman, of the Neurosciences Institute, New York, who postulates (eg. Edelman, 1994) what he calls a value system, the role of which he describes as follows: [//] 'What the value system does is it sends a chemical signal to the rest of the brain such that those connections that were just being used to produce [an] action which was valuable will become strengthened.' (Edelman, 1994, p11.)
You have only to touch a hot object to feel at first hand how well the system works!
In our own attempt at a three-layered control hierarchy (Smith, 1993), we take the ascending pleasure-pain pathway over to the right at the intermediate level, and then on up to the top level as an "antidromic" (or "backchannel") information flow [see pathways 2d and 2e on the Smith, 1993 diagram, and our e-paper on Antidromic Neurotransmission].