‘Artificial’, as the name suggests, refers to something which is man-made. Right from food flavoring to satellites, human beings have successfully created artificial versions of almost every natural thing. All these could be possible only because of one asset which human beings have endowed, ‘Intelligence’.
It has been well said that human imagination can travel faster that light. So fast that, it could even imagine the possibility of creating Artificial Intelligence (AI). Science fiction writers have used the possibility of intelligent machines to advance the fantasy of intelligent non-humans, as well as to make us think about our own human characteristics.
In the last half century, the AI community has been able to build experimental machines that test hypotheses about the mechanisms of thought and intelligent behavior and thereby demonstrate mechanisms that formerly existed only as theoretical possibilities. The history of AI is a history of fantasies, possibilities, demonstrations, and promise. Now let us start with an interesting question:
It has been well said that human imagination can travel faster that light. So fast that, it could even imagine the possibility of creating Artificial Intelligence (AI). Science fiction writers have used the possibility of intelligent machines to advance the fantasy of intelligent non-humans, as well as to make us think about our own human characteristics.In the last half century, the AI community has been able to build experimental machines that test hypotheses about the mechanisms of thought and intelligent behavior and thereby demonstrate mechanisms that formerly existed only as theoretical possibilities. The history of AI is a history of fantasies, possibilities, demonstrations, and promise. Now let us start with an interesting question:
“What is the difference between human intelligence and artificial intelligence?”
From a random web search, I came across this:
“Natural is your brain and artificial is google.com, but google.com was made by natural intelligence, in fact, all artificial intelligence comes from natural intelligence but no natural intelligence is from artificial intelligence.”
The difference between Natural Intelligence (e.g human intelligence) and Artificial Intelligence is hard to define, as not much is known about natural intelligence (NI). AI can learn, just like NI. When programmed to, AI can sense changes in its environment and react accordingly. It can then refer to the ways it reacted to previous changes to help decide what to do the next time a similar change occurs.
One big difference between Artificial Intelligence and Natural Intelligence is the fact that NI can forget and lose information. AI could do this only if it was program to do so, but this would be counter-productive. Another big difference is accuracy. AI, when given the same information can be exact, every time with speed. When NI is given the same information, it cannot be as exact, and is slower.
I discovered a couple of very interesting things few years back when I came across the computer software, ‘Dragon Naturally Speaking’. Dragon NaturallySpeaking is software that allows us to speak into a microphone device, and the computer will type out everything we say.
Whenever we talk to a new person our brain grasps certain characteristic features of that person which helps us to recognize that person, his face, his voice, in the future. A small talk is enough for it. But unlike the human brain, the artificially intelligent computer cannot do it so easily. When running naturally speaking for the first time, we need to make a user profile for ourselves. Then, we must train our voice with the new software. In other words, making the computer familiar with our voice so that it recognizes it next time, along with all the custom words we might have saved. This is done simply by reading some text into the microphone.
The most important thing in any speech recognition software is to speak clearly. The human mind knows what words are and how they are used, but a computer does not know what a word means. It cannot use assumption based skills to figure out what we are trying to say. For example, when I dictate the sentence, “I want to eat an ice-cream”, the computer might as well type, “I want to eat an I Scream”. (Also words like "four" and "for", and "red" and "read"). A human being is smart enough to figure out that ‘…I Scream’ is irrelevant with the phrase ‘I want to eat an…’. So it understands that it has to be ‘Ice Cream’ which is edible.
The most important thing in any speech recognition software is to speak clearly. The human mind knows what words are and how they are used, but a computer does not know what a word means. It cannot use assumption based skills to figure out what we are trying to say. For example, when I dictate the sentence, “I want to eat an ice-cream”, the computer might as well type, “I want to eat an I Scream”. (Also words like "four" and "for", and "red" and "read"). A human being is smart enough to figure out that ‘…I Scream’ is irrelevant with the phrase ‘I want to eat an…’. So it understands that it has to be ‘Ice Cream’ which is edible.
Can we expect the same thing from the computer?
Well, to some extent, yes. With repeated use of certain kind of words with certain phrases, the computer does learn from its mistakes. So the golden rule of dictation is to not our hands (in other words type) while we are dictating. There are a couple reasons for this; one is that we are robbing the software of an opportunity to learn from its mistakes. As we use the software and correct it, it learns and improves. This information is stored in our profile.
Considering all the pros and cons, I am actually amazed at how well it knows my voice. Granted, it does make some mistakes, but it is merely types what we say. And it does a great job at that. Kudos to speech recognition and artificial intelligence!
Saptarshi Prakash,
EE09B076,
References:
No comments:
Post a Comment