AL vs. AI
A friend's mother recently asked her, "How is pigseatingfigs' course in Artificial Insemination going?"
I thought that was rather good, and the friend said she had visions of me in a white labcoat splicing ostriches with horses, or somesuch. It would be rather more interesting than what I am doing at the moment, though.
Artificial Life is very easy to create. The seven qualities (I forget them, I hated Biology), include Feeding, Reproduction, Sensitivity and similar. They do not include intelligence.
This is an icon from an icon-criticising community:
It is Conway's game of Life, one of the first attempts to create Artificial Life. The black dots represent cells, and these live, die, or reproduce, depending on how many of the neighbouring cells are occupied. Sometimes these make stable forms, such as the 'fliers' or the group of 4 cells. One could say Conway was successful: these cells behave in a very similar way to bacteria, for example.
Artificial Life is very different to Artificial Intelligence, which is what I do. AI can be divided up into 'hard' or 'soft' AI. Overall, the aim of AI is to create intelligent systems (ones that can problem-solve intelligently). 'Hard' AI is what may be thought of as classical AI, like the film, the robots taking over the world, Lt. Data if you're a Trekkie, and so on. 'Soft' AI systems are systems which can solve problems, but they are not self-aware - they do not know they are intelligent.
When I started my course, I wanted to recreate hard AI. This is damn near impossible (now, and forever - if you would like to know why I can go into it at some length, but it is not all that interesting). So instead, I work on soft AI, and study Distributed Artificial Intelligence. This is when many, supposedly dumb (think of ants) beings work together to produce very efficient outcomes. Ants are not 'clever', but they build amazing and complex nests, they gather food efficiently, and so on, and they do this by working as a community, not as individuals. The fun thing about this is that very simple rules can lead to very complex behaviour. Look at birds - three basic rules (1. head for the middle of the flock, 2. fly in the general direction of the flock, 3. do not fly within 20cm of any other bird) can lead to beautiful flocking/flying behaviour.
It's interesting - it is a beautiful mathematical problem, but... I don't get to play God, so what is the point? I get very attached to my robots: I even believe they have personalities; but this is rubbish. Still, they're very pretty. Expect some screen shots once I get round to programming that part.
I thought that was rather good, and the friend said she had visions of me in a white labcoat splicing ostriches with horses, or somesuch. It would be rather more interesting than what I am doing at the moment, though.
Artificial Life is very easy to create. The seven qualities (I forget them, I hated Biology), include Feeding, Reproduction, Sensitivity and similar. They do not include intelligence.
This is an icon from an icon-criticising community:
It is Conway's game of Life, one of the first attempts to create Artificial Life. The black dots represent cells, and these live, die, or reproduce, depending on how many of the neighbouring cells are occupied. Sometimes these make stable forms, such as the 'fliers' or the group of 4 cells. One could say Conway was successful: these cells behave in a very similar way to bacteria, for example.
Artificial Life is very different to Artificial Intelligence, which is what I do. AI can be divided up into 'hard' or 'soft' AI. Overall, the aim of AI is to create intelligent systems (ones that can problem-solve intelligently). 'Hard' AI is what may be thought of as classical AI, like the film, the robots taking over the world, Lt. Data if you're a Trekkie, and so on. 'Soft' AI systems are systems which can solve problems, but they are not self-aware - they do not know they are intelligent.
When I started my course, I wanted to recreate hard AI. This is damn near impossible (now, and forever - if you would like to know why I can go into it at some length, but it is not all that interesting). So instead, I work on soft AI, and study Distributed Artificial Intelligence. This is when many, supposedly dumb (think of ants) beings work together to produce very efficient outcomes. Ants are not 'clever', but they build amazing and complex nests, they gather food efficiently, and so on, and they do this by working as a community, not as individuals. The fun thing about this is that very simple rules can lead to very complex behaviour. Look at birds - three basic rules (1. head for the middle of the flock, 2. fly in the general direction of the flock, 3. do not fly within 20cm of any other bird) can lead to beautiful flocking/flying behaviour.
It's interesting - it is a beautiful mathematical problem, but... I don't get to play God, so what is the point? I get very attached to my robots: I even believe they have personalities; but this is rubbish. Still, they're very pretty. Expect some screen shots once I get round to programming that part.