Knowledge Weight

When I think of AI I do not think of things like making a robot run or a program that can beat a human at chess. While these are difficult enough I consider them more mechanical in nature. 

My interest in AI has always been writing code that can emulate human thought. I do not believe we can ever write a program that thinks like a human because as a programmer I know a coder must understand the algorithm before he can program a computer to do it and to the best of my knowledge we still don't know the meaning of life so I don't know how we could write a program to live, but I digress. 

Part of the reason it is so difficult to model human thought is because we really don't quite understand everything around us yet and so we are not really sure how we actually accomplish a lot of what goes into the act of thinking. 

Just to be clear, in my opinion, humans think by manipulating concepts. In my world view everything is a concept. In fact our language, while conforming to various grammatical rules is typically spoken in idioms which seem to me to be a bit closer to the concept of a concept :-)

Now the problem with abstracting out knowledge stems in part from the issues we typically run into regarding reference. I have expounded on this in other writings in more detail, so just to summarize, a reference to a living human becomes a reference to a null object if the human is cremated. What should be left rather than a reference to a non existent human is a reference to a concept. Coupled with the concept of aliasing we can then deal with statements like "Samuel Clemmons was the writer known as Mark Twain" a bit easier.

But this alone still leaves us with issues like how do we even know that Samuel Clemmons actually existed? There is a belief among some that the combined works of William Shakepshere were not really an individual's writings but rather a collection of different individual's work over the years. In this way we can also deal with the concept of William Shakepshere, who in either case no longer walks among us. 

Which brings us to the topic of this post which is 'just how sure are we of what we think we know'? Now the process by which we accumulate knowledge is a far deeper discussion than I can go into here though it is obviously related but for now we will focus only on the strength of the knowledge we have. The knowledge weight.  

Consider how sure we are a person is alive when we believe to be in their presence. Let's call this 99%. We are almost positive (only because I don't believe in absolutes) the person before us is our uncle, for example. Now compare that with how sure we are we that person is alive five minutes after we just got off the phone with them versus someone we haven't spoken to in six years. So as we can see, there are varying weights we can give to knowledge based on various other factors. 

By the way, another reason why I say 99% and never 100% is because the person before us could be an impostor. They could be someone who looks just like the person we think we know. We could be fooled. If we didn't interact with them, and perhaps just passed them on the street we could be less sure. So knowledge may have weight. But there are classes of knowledge so maybe it would be better to categorize knowledge and then add a weight to that. 

For example, I think I know my friend Bob because when I am in his presence I can see, feel and  hear him. I recognize him from a combination of attributes and mannerisms. I also think I know that someone named Ghengis Khan existed at some point in time. They are two different categories of knowledge; direct and indirect. Anything my father told me is indirect knowledge. Anything I have physically interacted with is of a different category and within these categories I can assign a weight.