Artificial Intelligence
Tuesday, October 19, 2010
Objections to Artificial Intelligence
One of the most famous objections to artificial intelligence is John Searle's Chinese Room argument. This argument by Searle offers up an idea that a machine can do a certain task, but does it actually have an understanding of the task. This is a great question in the field of AI. Sure a machine can do a certain task, but does it grasp and have an understanding of what it is actually doing? This poses questions whether or not AI can actually model human processing and consciousness.
Here is a link describing Searle's Chinese Room argument.
http://plato.stanford.edu/entries/chinese-room/#3
Another one of the most famous objections is the Turing test. Many people use the Turing test as a way to test if a computer can actually think. In the Truing test there is a human judge that has a conversation with a human and a computer (they're all in separate locations). It is the judges job to figure out with is a computer and which is a human. They do this test through text-only conversation through a computer screen and keyboard. If the machine can pass as a human, it has passed the Turing test.
Here is a link to a description of the Turing test.
http://en.wikipedia.org/wiki/Turing_test
Another famous objection is one given by Lady Lovelace. Lovelace's objection with AI is that it cannot have any original ideas, or free will. A computer can merely do whatever information and rules it is given. If computers have no free will, how can they grow as programs?
Another objection is that programs in artificial intelligence that stems from Lady Lovelace's objection is that computers are only good in their specific domain. If they are given input from a different domain where they have no way of processing the information, the program is considered useless.
There is no right or wrong answer when it comes to the field of artificial intelligence. There are many people who believe many things about artificial intelligence. Some promoting and believing that artificial intelligence can be successful, and others who believe that AI cannot model human processing, and that it never will. It is up to you to draw your own conclusions about artificial intelligence. As long as there is technology and great want to understand the workings of the human mind, there will be great debates about the field of artificial intelligence and its applications.
Here is a great site about AI that contains good objections about AI and also replies to those objections.
http://www.iep.utm.edu/art-inte/#H4
Another site containing objections to artificial intelligence.
http://www.xuenay.net/objections.html
Neural Networks
The human brain is made up of millions and millions of cells called neurons. These neurons are connected to other neurons, creating a large "network" of interconnected neurons that create the central and peripheral nervous systems.
In the field of artificial intelligence, researchers study the workings of biological neurons through the use of artificial (replica) neural networks. These programs mimic the properties of real neural networks in hopes to gain a better understanding of how biological neural networks actually work.
These simple models of neural processing have helped gain lots of information and insight on how the human brain might work with respect to neural processing. Neural networks look at how the brain is wired through networks of neurons and synapses, and draw conclusions from this information. Believe it or not, neural networks can also learn just like you and me!
Neural networks can learn from positive, and negative reinforcement just like humans can. Though it does happen at a level that isn't as high as humans, it is still insightful and can help model certain processes.
There are also many other applications of neural networks besides using them for furthering knowledge on the human mind. There are applications of neural networks in business, medicine, robotics, and credit scoring to name a few. Who would've thought that a neural network could have a vast array of applications?
Neural networks are very important to the field of artificial intelligence as well as many others. Neural networks are such a large and complicated subject, that a mere blog can't even begin to scratch the surface on this amazing field. However, I've posted some links to help you hopefully better understand what neural networks are, and how they function.
For next time, we'll talk about the many objections to artificial intelligence.
Here is a bank of information about neural networks.
http://www.doc.ic.ac.uk/~nd/surprise_96/journal/vol4/cs11/report.html
Here is a car that is driven by neural networks
http://www.youtube.com/watch?v=Wn6nRaPc_Pw
http://en.wikipedia.org/wiki/Hebbian_theory
Multilayer Perceptron Neural Network
http://www.dtreg.com/mlfn.htm
Who can ever get sick of watching RoboCup
http://www.youtube.com/watch?v=iMM_XQXJUUc