Artificial Intelligence - Technical Nuances - Part 2
Read Part 1 - Artificial Intelligence - A New Future - Part 1
In my last post on AI, I had ended saying it was a new future indeed! Amazing wasn’t’ it? That AI could do so much, bring so much change, that too at a massive scale!
If I have to explain AI in short then I will say, "AI is the technology which helps, computers to think and behave like humans". And this is validated by the Turing test. The Turing test was created by Alan Turing to determine whether a machine is intelligent. How does the test work? A human will ask questions to another human and a machine and based on the answers, the questioner will decide who is the human and who is the machine. If the machine manages to convince that the questioner that it is a human, then we can say that the machine is "Artificially Intelligent". Oxford dictionary defines intelligence as follows, "The ability to learn, understand and think in a logical way about things". Here, it becomes critical for me to mention that AI is a broad concept, AI has a subset which is Machine Learning, and Machine Learning is the ability of a machine to learn on its own. And this is important, as, without learning, there is no intelligence.
If I have to explain AI in short then I will say, "AI is the technology which helps, computers to think and behave like humans". And this is validated by the Turing test. The Turing test was created by Alan Turing to determine whether a machine is intelligent. How does the test work? A human will ask questions to another human and a machine and based on the answers, the questioner will decide who is the human and who is the machine. If the machine manages to convince that the questioner that it is a human, then we can say that the machine is "Artificially Intelligent". Oxford dictionary defines intelligence as follows, "The ability to learn, understand and think in a logical way about things". Here, it becomes critical for me to mention that AI is a broad concept, AI has a subset which is Machine Learning, and Machine Learning is the ability of a machine to learn on its own. And this is important, as, without learning, there is no intelligence.
So how do these machines learn on their own? Just like us!
By collecting data through seeing (cameras), hearing (microphones), sensing (sensors)! Therefore, they learn by analyzing vast amounts of data which are collected through multiple ways like cameras, microphones, LIDAR based sensors, Natural language processing, etc. LIDAR is an acronym for light detection and ranging which is the technology through which autonomous vehicles like driver-less cars function.
Data is food to AI. Big data is a combination of structured and unstructured data. AI analyses data with the help of neural networks, data mining algorithms and tries to form patterns and connections to come to a conclusion. (I hope I succeeded in explaining this without much technical jargon or anything Please leave comments if I didn't)!
There are two types of learning i.e. Supervised - Learning with the teacher and Unsupervised – Learning without the teacher. In supervised learning, the output datasets called training data sets are used to train the machine to get the desired outputs. A real-life example would be of classifying whether a patient has a disease or not, based on symptoms. The computer is fed the data of symptoms and it is given a disease name for each data set of symptoms, and hence when you ask the system for diagnosis, it will try to correlate your set of symptoms with the disease labels.
Whereas in unsupervised learning no training datasets are provided, instead the data is clustered into different classes. For example, you give the machine, the data of customers' purchases, their background, age, gender, ethnicity, income, etc. The machine will try to correlate the data and try to form patterns and classify it by way of clustering i.e. – Men working in Law prefer to buy white shirts more than men working in Advertising; or by way of association - Children’s of age 3 to 7 buy Doremon Bags, while children of 7 to 11 will buy Chotta Bheem bags, and when they buy bags, they also buy bottles in a similar fashion. Here the computer associates bags with bottles.
Many recent innovations are on the middle ground i.e. semi-supervised learning, wherein you have a large data set, but it is impossible to label or classify all the data. So, you label a small set of data and then the other data is classified by the machines, based on its training set.
This completes the technical part of AI.
More on AI in my next post. I will be writing about the Impact of Artificial Intelligence on employment.
This completes the technical part of AI.
More on AI in my next post. I will be writing about the Impact of Artificial Intelligence on employment.
x
Comments
Post a Comment