What is Artificial Intelligence

What is Intelligence

Albert Einstein, one of the most astounding geniuses the world has ever seen, had an intelligence quotient of 160. But is genius same as intelligence? Definitely not.

Even though experts do not agree on a single definition for the term intelligence, generally, intelligence is regarded as the cognitive problem-solving skill (Miller, 2010). It has to do with the mental ability that is required for perceiving a relationship, reasoning, learning at a faster rate, calculating, among other things. Psychologists used to believe that there was a single factor, referred to as the g-factor, which was responsible for one’s intelligence. This belief was later on cancelled out after it was realized that intelligence is more complicated and that such a simplistic method could not be the basis for citing one’s intelligence.

Different psychologists have come up with their own categorization of intelligence. For instance, the reputable Howard Gardner developed the seven components that ought to be looked at when discussing the subject of intelligence. These components are: linguistic, bodily-kinesthetic, musical, intrapersonal, spatial, interpersonal and logical-mathematical.

An individual’s ability when it comes to the understanding of complex ideas, learning from experience, effectively adapting to a new environment, and overcoming obstacles through detailed thinking, differs from that of other people. As much as these individual differences are substantial, it is not possible to establish some consistencies. That is to say that your intelligent performance will vary according to the situation in which you are.

Whatever definition is adopted for the term intelligence, the baseline is that the definitions are all focused on certain behaviors. It is primarily involved in how well you are able to comprehend your surroundings, figure out what to do, and make sense of things.

So what is AI

Whereas intelligence tend to be focused on humans and other living creatures as a whole, artificial intelligence concentrates on machines.

Artificial intelligence, abbreviated as AI, is defined as the ability of a computer-controlled robot or a digital computer to carry out different tasks that are typically linked to the intelligence of a human being (Kaplan, 2016). In simpler terms, it is the ability of a machine to do things that a normal human being is able to accomplish.

When you are developing a system that is supposed to do things “like a human being,” assessing the AI of that system will be focused on whether it can reason, learn from experience, discover meanings, and generalize. Ever since digital computers came to existence in the 1940s, different programmers have shown that we can actually program a computer to take some behaviors of a human being. For instance, such programmed computers could be used to play chess or discover complex mathematical theorems.

Sophia is one of the most comprehensive AI robot. Developed by Hanson Robotics which is based in Hong-Kong, Sophia became active (alive if you like it) on 14th February, 2016. That’s basically the robot’s birthday date! Sophia became the first non-human to receive a United Nations title when the robot was named the UN Development Programme’s Innovation Champion. The AI robot, can do several things including recognizing individuals, following faces and sustaining eye contact. This is because of the cameras fitted onto the robot. These cameras are controlled by a set of algorithms. Sophia is able to process speech and even hold a conversation via a natural language subsystem. Recently, Sophia was upgraded after being given functional legs and the ability to walk (Jaden & MacKenzie, 2018).

All these things that Sophia is able to do is basically the culmination of what artificial intelligence is all about. Because the things that machines can do continue to expand, there is a dispute when it comes to the scope of AI. AI effect is a concept that emerging from all these disputes over what artificial intelligence is actually all about. It is captured in Tesler’s Theorem which holds that AI is that thing which is yet to be done. For instance, optical character recognition has become a routine technology such that it is no longer included in the definition.

Various machine capabilities like understanding human speech, playing games that require high-level intelligence like chess, intelligent routing, and autonomously operating cars are all classic examples of artificial intelligence.

It must be understood that AI was established on the basis of human intelligence (Wang & Goertzel, 2012). During its earlier days of development, researchers were largely guided by the claim that it is possible to describe human behavior in such a manner that this behavior is simulated. That has always raised philosophical arguments with regards to the nature of the mind and the ethical implications of developing machines that are able to ‘act as humans.’ Given that these machines have human-like intelligence, there are those who strongly believe that AI is a danger to humanity, besides increasing cases of unemployment (Türk & Savacı, 2006).

Strong AI and Weak AI

Weak AI, also referred to as narrow AI, is basically artificial intelligence that concentrates on narrow task. Strong AI on the other hand is the actual definition of artificial intelligence as we know it (full AI) (Kerns, 2017). Academic sources reserve the definition of strong AI for machines that are able to experience consciousness.

Weak AI has been around for quite a while, therefore it is a tested field with continual developments to further improve it. Computer games is one area where weak AI has been strongly applied. In this case, a person competitively plays a game against a computer without realizing that it’s actually a computer that’s playing. This is because the algorithms written for such games are so detailed that they are able to make the computer mimic human behavior (Human Paragon, 2017).

Engineers and companies that require robots to perform certain tasks typically use this concept of weak AI. The tasks could be as simple as picking an item and placing it on a belt. However, given that the tasks has to be done repeatedly and for long hours, sometimes it becomes inconvenient to use human power.

We may consider Alexa and Siri to be AI, however, these are all part of weak AI programs. Even the most chess is weak AI. The classification is basically based on the concept of supervised and unsupervised programming. Chess programs and voice assistance are typically fitted with programmed response. Their working is based on determination of keywords that they already know and then respond as already commanded. Even though it gives a human-like experience, it remains to be a simulation. For example, you may tell Alexa to turn on the music. The voice-assistant technology will pick out key words like ‘on’ and ‘music,’ thereby turning on the music as programmed. Even though it has done what you wanted, it still does not comprehend the actions it just performed like a human brain would.

Strong artificial intelligence is the latest development in the field of AI. It is based on the human mind and attempts to equip machines with the ability of doing human duties without external help. The focus is that technologies ought to be developed to the extent where they no longer depend on human beings. Even though this is a new field, Sophia clearly demonstrates that it is a field which is rapidly developing. As more details are obtained from research, we could soon have a community of machines that no longer need our input to do the tasks they are programmed to perform. For instance, they could ‘wake up’ early to prepare you breakfast long before your alarm goes off (Kerns, 2017).

Typically featured in movies, strong AI is an exact mimic of the brain. Rather than classification, it processes data via clustering and association. In simpler terms, this means that there are no predefined answers for given keywords. The function is able to mimic the result but there is no certainty as to what the result would. It is similar to talking to a human being. Even though you can make assumptions as to how an individual will reply to your question, you do not know the exact response. For instance, a machine would “hear” the term good morning and begin associating it with a coffee maker.

The popularity of automation and robotics in factories and homes has increased. The response of these tools tend to take a particular path depending on the input. One may be glad that they are able to play game with the machine or speak to a given robot, but syntax cannot suffice semantics. Whereas weak Ai has been instrumental in helping us enjoy the current state of robots, strong AI makes the robot listen to you.

Expert Systems

An expert system refers to a computer program which implements AI technologies by simulating human behavior and judgment. The main aim of expert systems is to emulate the decision-making ability of humans.

These systems normally incorporate knowledge base that comprises of accumulated experience and rules engine. Rules engine refers to directives that are followed in applying knowledge base to given situations that are described in the program. As more additions are made to the knowledge base of rules engine, the ability of the given system is enhanced.

There are some key characteristics associated with expert systems including: reliability; high responsive; understandable; and high performance (Techopedia, n.d.). Expert systems are capable of an array of tasks including: deriving a solution; advising; diagnosing; interpreting input; justifying the conclusion; predicting results; and explaining. At the same time, these systems are incapable of: processing human capabilities; refining their own knowledge; and substituting human decision makers.

The expert systems are primarily composed of three components, namely: knowledge base, inference engine, and user interface.

The knowledge base comprises of domain-specific and high quality knowledge. This is needed for the purpose of exhibiting knowledge. In order for the ES to be successful, it must have a detailed collection of precise and highly accurate knowledge (Techopedia, n.d.). Knowledge is basically the collection of facts. The expert system knowledge base stores both factual knowledge (information that knowledge engineers and scholars widely accept) and heuristic knowledge (associated with guessing, accurate judgment, and the ability of an individual to assess situations).

Inference engine has to do with the rules and procedures that the Inference Engine uses in coming up with a flawless solution. The Inference Engine usually capture knowledge from the knowledge base and use it in arriving at given solutions. There are two strategies that the Inference Engine follows, namely: forward chaining and backward chaining.

User interface is the platform that the user relies on while interacting with the expert system. It is built in a natural language to make it user-friendly. One does not have to be an expert in artificial intelligence to take advantage of the benefits of expert systems.

Even though expert systems have proved themselves useful in various domains, they also have limitations. Some of these include:

  • Difficulty in maintaining the ES
  • High development costs involved
  • Difficult knowledge acquisition
  • Limitations of the technology

Leave a Reply

Your email address will not be published. Required fields are marked *