Nicole (![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png)
trickykitty) wrote2008-01-24 12:03 am
trickykitty) wrote2008-01-24 12:03 am
Thesis Topic
I'll probably put up periodic updates on my senior thesis, just because I think it's a fun topic and I know some people would like to read it. I'll always hide it behind a cut.
Also, do keep in mind that I'm a horrible writer, so if you want to critique on the basis of grammar and such feel free, but be careful about critiquing the content. If you have content input just give me forewarning so I'm not taking personal offense to it.
Today's post is my thesis topic. It's only supposed to be a few paragraphs more or less summarizing what I will be writing about. Considering I just typed it up in 30 minutes flat, I'm proud that I was able to be as succinct as I was. I otherwise tend to ramble way too much about my favorite topic. Oh, and I will be turning this in first thing in the morning, so grammar be damned.
Artificial intelligence (AI) is a relatively new field of study which has broad applications in a wide range of divergent fields. There are two major approaches to AI as described by Alan Turing. These are known as the “bottom up” and the “top down” approaches.
The bottom-up approach is biological in nature and involves piecing together different types of neurons into a particular configuration and “turning them on” in an effort to simulate the human brain. The idea is that the neurons themselves will know what to do automatically and will begin doing their inherent job of creating a living, thinking brain. In this approach, it is important to know ahead of time what the neurons will be doing and it will be required to know how they do it. This is a slow process as millions of neurons are grown in biology labs, and it is restricted by our knowledge of neurons, neurotransmitters, and other aspects of the biological makeup of the brain.
The bottom-down approach, also known as the black box approach, is the version which most people think of when they hear terminology associated with artificial intelligence. The end goal is to mimic human (or animal) behavior without constricting the internal workings of the entity. An artificially intelligent machine has not been imbued with neurons, but instead has been programmed to mimic the actions and learning processes of the brain. Programmers gain knowledge through trial and error and gradually increase the functionality of the programs that are used. Just as the bottom-up approach is limited by biological science, so too is the top-down approach limited, in this case by knowledge of programmability as well as software and hardware constraints.
There is emerging a third area which combines these two approaches in an attempt to bridge the gap. This third approach involves neural networks, mathematical systems used to explain the interactions of nodes. Nodes are described, in this case, as brain areas which can be as small as just a few neurons or as large as a brain region (such as Broca’s area or Wernicke’s area) or an entire brain structure (such as the hypothalamus or the cerebellum). These mathematical systems are then translated into programs in order to simulate brain reactions in an effort that results in similar behavioral outcomes. Neural networks benefit from not having to specify each individual neuron’s response as required in the bottom-up approach, while also providing a clear framework for programming in the top-down arena.
The two main approaches have great value to society with regard to implementation within the fields that they incorporate. The biological approach has close ties with stem cell research and other biological studies working toward cures for disease and replacement body parts. “Intelligent” programs are used extensively in data mining, on vehicles, and even in doctor’s offices. There is inherent value in continuing both the bottom-up and the top-down approaches.
The question remains of the value of creating a truly artificially intelligent machine as these two areas of research continue to align. If a machine’s program is based on the real processes of the human mind, then there is the strong possibility that the machine may not choose to proceed as it was prescribed, just as a child rebels against its parents and does not act in accordance with the parent’s wishes.
The purpose of this paper is to provide an analysis on the value of artificial intelligence. Focus will be given to the two primary fields specifically, looking at current research, advances, and practical uses already in place, while also investigating theories of the future of AI research from the perspective of these merging spheres.
Also, do keep in mind that I'm a horrible writer, so if you want to critique on the basis of grammar and such feel free, but be careful about critiquing the content. If you have content input just give me forewarning so I'm not taking personal offense to it.
Today's post is my thesis topic. It's only supposed to be a few paragraphs more or less summarizing what I will be writing about. Considering I just typed it up in 30 minutes flat, I'm proud that I was able to be as succinct as I was. I otherwise tend to ramble way too much about my favorite topic. Oh, and I will be turning this in first thing in the morning, so grammar be damned.
Artificial intelligence (AI) is a relatively new field of study which has broad applications in a wide range of divergent fields. There are two major approaches to AI as described by Alan Turing. These are known as the “bottom up” and the “top down” approaches.
The bottom-up approach is biological in nature and involves piecing together different types of neurons into a particular configuration and “turning them on” in an effort to simulate the human brain. The idea is that the neurons themselves will know what to do automatically and will begin doing their inherent job of creating a living, thinking brain. In this approach, it is important to know ahead of time what the neurons will be doing and it will be required to know how they do it. This is a slow process as millions of neurons are grown in biology labs, and it is restricted by our knowledge of neurons, neurotransmitters, and other aspects of the biological makeup of the brain.
The bottom-down approach, also known as the black box approach, is the version which most people think of when they hear terminology associated with artificial intelligence. The end goal is to mimic human (or animal) behavior without constricting the internal workings of the entity. An artificially intelligent machine has not been imbued with neurons, but instead has been programmed to mimic the actions and learning processes of the brain. Programmers gain knowledge through trial and error and gradually increase the functionality of the programs that are used. Just as the bottom-up approach is limited by biological science, so too is the top-down approach limited, in this case by knowledge of programmability as well as software and hardware constraints.
There is emerging a third area which combines these two approaches in an attempt to bridge the gap. This third approach involves neural networks, mathematical systems used to explain the interactions of nodes. Nodes are described, in this case, as brain areas which can be as small as just a few neurons or as large as a brain region (such as Broca’s area or Wernicke’s area) or an entire brain structure (such as the hypothalamus or the cerebellum). These mathematical systems are then translated into programs in order to simulate brain reactions in an effort that results in similar behavioral outcomes. Neural networks benefit from not having to specify each individual neuron’s response as required in the bottom-up approach, while also providing a clear framework for programming in the top-down arena.
The two main approaches have great value to society with regard to implementation within the fields that they incorporate. The biological approach has close ties with stem cell research and other biological studies working toward cures for disease and replacement body parts. “Intelligent” programs are used extensively in data mining, on vehicles, and even in doctor’s offices. There is inherent value in continuing both the bottom-up and the top-down approaches.
The question remains of the value of creating a truly artificially intelligent machine as these two areas of research continue to align. If a machine’s program is based on the real processes of the human mind, then there is the strong possibility that the machine may not choose to proceed as it was prescribed, just as a child rebels against its parents and does not act in accordance with the parent’s wishes.
The purpose of this paper is to provide an analysis on the value of artificial intelligence. Focus will be given to the two primary fields specifically, looking at current research, advances, and practical uses already in place, while also investigating theories of the future of AI research from the perspective of these merging spheres.
no subject