Design Perceptive

There is no singular, unanimously accepted definition of artificial intelligence.  This is to be expected since there is no singular, unanimously accepted definition of intelligence.  In fact, most books about these topics, when dealing with the task of defining them, make this very assertion.  They then go on to quote multiple scientists and revolutionaries who took a stand on their opinions.  To be careful, the authors of these books cover the range of opinions, then they settle on one opinion because it is necessary to the purpose of the book.

For the sake of brevity, I’m not going to go into a wide range of opinions. (For a wide range, see the Wikipedia post for Artificial intelligence or go directly to Russell, Stuart J.; Norvig, Peter (2003), Artificial Intelligence: A Modern Approach (2nd ed.), Upper Saddle River, NJ: Prentice Hall, ISBN 0-13-790395-2 )  Instead, I’m going to tell you what I think and which authors and scientists have influenced my own opinion.

First, intelligence is an organism’s ability to learn about, respond to and predict about changes in its environment (See Hawkins, J. On Intelligence, Time Books, 2004.)

Now, I’m sure I’ve already upset some people.  I already feel the thousands of objections and questions beating down my door.  This is also to be expected.  These are bold, wide-reaching statements.  Many organisms are covered under this concept.  Intelligence is not the exclusive property of Man (or Beast).  Intelligence doesn’t come about strictly because of the neo-cortex.  Yes, Dorothy, things aren’t black and white.  There are shades of gray in this universe and intelligence is one of them.

In On Intelligence, Jeff Hawkins describes a memory-prediction framework for describing what we, as humans, generally call thinking.  In short, the main idea is that bottom-up sensory inputs are matched against stored patterns which, in turn, provide a top-down prediction in the form of neural potentiation. ( http://en.wikipedia.org/wiki/Memory-prediction_framework )  The book goes into great depth about this framework, provides a basis for thinking about how the brain works and answers a great many questions.

In following this model, we accept that there are three parts to intelligence:

1. Ability to learn
2. Ability to respond to
3. Ability to predict

If we accept this, then we must also accept that intelligence exists on a continuum. 

I’ll let you digest this for now.  In the next post, we’ll go into why I think this way, what some objections to these concepts are and what the implications of this concept are for computer science and the field of artificial intelligence.

Temporal Locality, in regards to computer memory management, says that if an item is used, it is probable that it will be used again soon.  It is part of a principle in computer science called the Locality principle.  This principle maintains that data locations referenced in a short period of time in a running computer often consist of relatively well-predictible clusters¹.  It was developed as the theory behind a robust page replacement algorithm for virtual memory in the 60′s.  Since then, the principle has been applied in a number of ways.  I think we can use it in another, useful and interesting way.

We might apply this principle in vocabulary selection for speaker-independent speech recognition systems.  Because size and composition impacts speech recognition performance, it is important, at any given time in a conversation, to maintain a reasonable vocabulary size.  Yet, we want vocabularies that are relevant to the context.  Obviously, we can code our own vocabularies and vocabulary swaps like the programmers of yesterday who had to manage their own page transfers.   But, this is prone to human error and does not rely on the power of computation.  How can we automate vocabulary selection and swapping?

Denning gives us the hint in his article in Communications of the ACM “The Locality Principle”.  It is useful to provide the entire quote here.

The locality principle flows from human cognitive and coordinative behavior.  The mind focuses on a small part of the sensory field and can work most quickly  on the objects of its attention… …The locality principle will be useful wherever there is an advantage in reducing the apparent distance from a process to the objects it accesses.

So, it seems to me, and this is my hypothesis, that the vocabulary of human discourse follows the same principle.  Simply put, working vocabularies have a reference of locality with the vocabulary domain of the discourse context.  This is something I intend to explore over the next year.

1. Denning, P.  ”The Locality Principle” Communications of the ACM, July 2005 [↩]