Intelligent home philosophy

It's hard to deny that simply wanting to investigate intelligence in a very structured way is not at least part of the driving force for this extended work. However, there is also the conviction that Machine Intelligence has the potential to be a very useful and desirable proposition even if it is really going to take some doing to actually bring it about, make a truly intelligent machine that is. Make the intelligent machine without simply changing what intelligence means or using terms like 'smart' instead. There are working definitions for machine intelligence but they don't match up with the ones for intelligent humans and they in turn don't match up with the ideas about what an intelligent dog can do. The Turing test is an admirable goal but to satisfy its conditions in isolation will not necessarily lead to an intelligent machine ,as many chat bots, really good chat bots, are proving.

Machine Intelligence will need to be embedded in a Machine based entity, something which can have it's own existence, carry out its own activities and sense its own environment in its own way. We might end up referring to the intelligent bit as a set of software systems at the top, but without the embedded nature of these systems, there will be no intelligence. That is, we will not make an intelligent computer which does not have some sort of body, even if that body is a very different sort of body to the ones we are used to.

Intelligence embedded in an Entity


This Intelligent Home Project is intended to be a great way to think about things. Thinking is quite a good thing to do for the most part. Any machine intelligence which may emerge from this project will emerge within a system which has a body and a place to live even if the place to live is itself somewhere where people live. This Intelligent Home will have a brain which has fast response bits, memory for things that have happened, ways to work out what might happen next, ways to decide what to do next and many other things which a brain needs.

If the intelligence is ever able to converse, it will have things to talk about based on what it knows about the world. However, the way in which it will carry out any conversation will depend on what it can do based on its own capabilities, sensing systems and bandwidth and its own output capabilities. It will be a machine which survives in its own niche and its capabilities will be developed for that niche.

An intelligence like human intelligence, which can learn from conversation and can form abstract concepts is a truly exciting goal. However, a machine which can do this should be able to do it as well as some more fundamental intelligent things such as working out what to do next in a new situation and not instead of them.

What is the point of this?

Intelligence, like knowledge and like wisdom, are concepts with considerable scope to explore. In other words, there are those who may find the exploration of intelligence, knowledge and wisdom, quite interesting. Some of the problems that I (from my personal perspective) have with AI (Artificial Intelligence) is that it seems to get decomposed into very small systems to such an extent that it is really hard to tell why the system or idea being explored isn't simply computer software. Once you take the bits of AI out of the context of an intelligent machine and then turn the bit into a branch of AI in its own right, it really stops being AI, in my opinion. That is not to say that it stops being useful, it just stops being AI. I can look at many (if not all) of the (AI) books in my own small library of books and few if any of them actually mention a fully intelligent machine, even the idea that the things in the book are intended to be part of one. The part of my book collection concerned with mind and brain comes closer to discussing intelligence than the AI books.

Whilst this Intelligent Home project may not achieve a great deal, it will be an embedded and integrated project which at least has the potential of developing an intelligent machine. That is really the point, to work on AI (or MI) in such a way as to be able to show what contributions each of the separate bits are making to the overall and developing intelligence. Any successful bits can be shown within context even if they are independently useful. From the earlier research, the redesigned AMC has a very important part to play within the context of overall intelligence and yet it is quite useful as an independent component. There is nothing wrong with the bits being independently useful but they need a whole system context of the Intelligent Machine if they are to be considered as part of an investigation into AI. If the bit only has an independent use, and the use does not create an intelligent machine, then it is not really AI.

There are contact details above if this view really does strike a nerve.