Artificial Intelligence is one of the most trending topics of the 21st century, and yet it is also amongst the most debated upon topics. People fail to understand the basic concepts of AI and frequently asked questions include, “Will machines dominate the human race?" or "Will AI be a threat to humans as it can lead to mass unemployment".
This article aims at relieving the general public of misconceptions with respect to AI and making them aware of the important roles that AI will be playing henceforth in their lives.
We'll start with what Alan Turing in the 1950s gave as the most fundamental definitions of Artificial Intelligence. "Artificial Intelligence is a way of making a computer to think intelligently, in the manner the intelligent humans think". It is all about machines being able to think and act as intelligent humans. Accordingly, broadly AI is classified into - strong AI and weak AI. The work aimed at genuinely stimulating human reasoning is called as "strong AI". As such a system is difficult to comprehend as our readers might have guessed, no such real model of strong AI has been developed so far. Such a realisation would have led to an actualisation of the movie "Terminator Salvation".
The other category namely “weak AI” on the other hand has seen numerous developments as it aims at simply getting the work done, it being done in a non-human way altogether. An excellent example of such a system is IBM's Deep Blue that is a master chess player, but it does not play in a way similar to humans.
We also have a third categorisation somewhere in-between the two. IBM's Watson is such an example. It compares a given text with its huge collection of database to arrive at the conclusion of a pattern in it, similar to what humans do on the form of image processing.
So how do we develop AI? That is the question. As with all other applications, this too involves programming. But this is a lot different from structural programming. A program without AI can answer the "specific" questions it is meant to answer, whereas that with AI answers any question belonging to its "generic" type. If you modify a structural program, its entire structure changes. AI programs are all about modifications. They keep absorbing the information provided to them as stimuli for future referencing like the human brain.
What is intelligence? Are you intelligent? Let me tell you folks, you are not intelligent if you don't know what it means. Intelligence is intangible and is composed of - reasoning, learning, problem solving, perception and linguistic knowledge. Let’s continue our discussion.
Let’s discuss some basic AI terminologies -
1. Environment and agents : An agent is anything that can perceive its surroundings through sensors and acts upon that environment using effectors.
-> A human agent has the 5 sensory organs for receiving stimuli and the motor neurons attached to other body parts such as the appendages to act as effectors.
->A robotic agent on the other hand has cameras, ultrasonic sensors, infrared sensors etc. to receive stimuli and motors and actuators as the effectors.
->A software agent has encoded bit strings as its programs, user inputs as stimuli and effectors such as printers, speakers, etc.
2. Search Terminologies :
-> Problem space : The environment in which the search takes place.
-> Problem instance : It is the initial state(given) + the goal state(mission).
-> Problem Space graph : It represents problem state. It is a graph showing states as nodes and edges as operations.
-> Depth of a problem : There can be many paths to solve a problem. The length of the shortest path from the initial to the goal state(number of edges) is called its depth
-> Space complexity : The maximum number of nodes to be stored in the memory.
-> Time complexity : The maximum number of nodes that are created.
-> Branching Factor : The average number of child nodes in the problem space graph.
All these are the formalities involved in solving a simple problem like solving a 3X3 - 8 tile problem (check cover photo).
I would like to conclude this article with the infamous Alan Turing Test. Turing Test is an annual battle between the world’s most advanced artificial-intelligence programs and ordinary people. The objective? To find out whether a computer can act “more human” than a person. For the past 2 decades the AI community has been convening every year for its most anticipated event, the Turing Test. The winner of the Turing Test gets the Loebner Prize with a prize money of $3000. Several judges each pose questions, via computer terminal, to several pairs of unseen correspondents, one a human “confederate,” the other a computer program, and attempt to discern which is which. The dialogue can range from small talk to trivia questions, from celebrity gossip to heavy-duty philosophy—the whole gamut of human conversation. The judges have to decide out of each pair, which one is the human. Failing to do so makes the computer program the winner. Sounds interesting? Here is a link to the experience of one such human confederate Brian Christian :
That’s all for now folks!