Artificial Intelligence (AI): The Intelligence Exhibited By Machines.

This Wikipedia definition of Artificial Intelligence is shrouded by many inherent questions. One of such is; how can machines exhibit intelligence? Researchers have always attributed the human subconscious ability to be intelligent to implicit mathematical calculations occurring in the brain. If this is true, then the mathematics of human intelligence can be modeled and applied to creating machines that were truly intelligent. This thought process led to the question, Can machines think? posed by the famous computer scientist Alan Turing in his paper “Computing Machinery and Intelligence” in 1950. Artificial Intelligence (AI) can thus be redefined as a science aiming to achieve the creation of machines and software that can think; Intelligent.

 

“Nkechi is Intelligent; she knows what to do, when to do it and how to do it”. This scenario is a typical human scenery composed of intelligence. It thus implies that a truly intelligent machine would be able to perceive its environment and act in order to achieve a goal whether simple or complex. This means that an intrinsic component of Artificial Intelligence is the ability to Perceive. How do humans perceive? Sight, Hearing, Touch, Smell and Taste. Can we build machines that can have all or some of these senses?

 

Another component of Artificial Intelligence is the ability to interact with a perceived world i.e Walk, Talk, etc. Can we build machines that can do all or some of these? Also, another implicit component is the ability to learn. Can we build machines that can learn from experiences and take decisions based on what they have learnt? The answer to most of these questions is Yes.

ai4

 

In 2012, Andrew Ng at Google, trained a machine learning algorithm to watch YouTube videos and recognize things. For one week, the algorithm went through thousands of videos and when it was done, it could recognize 12 thousand objects including animals and human beings. This shows that there exist algorithms that already have the ability to see and recognize just by training and learning. The algorithm performed well with over 90 percent recall; better than some humans. The group of algorithms by which an algorithm is taught to learn is called Machine Learning.

 

Using Machine Learning, Google DeepMind led by Demis Hassabis created an AI that could play the game Go. Go is said to be most complicated game in the world because the possibility of moves on a 19 by 19 board is intractable. In fact, it is 20 raised to the power of 170. This AI, Alpha GO learnt to play this deeply complicated game and in a series of games against world champion Lee Sedol last month, beat him. This was referred to as an important day in the history of AI research.

 

Of course it is common knowledge that there are cars that drive themselves. They see the road, cars ahead, pedestrians and making informed decisions on how to drive. There also are algorithms that talk e.g SIRI and Facebook. On the local level, an affiliate of Courteville Business Solutions Plc, Olumide Okubadejo majors in Computer vision and when simply put it means giving intelligent eyes to algorithms and Machine Learning which is teaching machines to learn worked on an algorithm at the University of Southampton that can see how people are walking and recognize them by their walking steps. He is also currently working on how algorithms can see things and differentiate the tiny details even to the difference between two grains of sand.

 

It would be noticed that there is an individualistic approach to creating AI, an approach where sight is treated independently from the ability to hear. This is based on the inability of research to scale without incurring computational costs. Although, AI research have taken giant strides in individualistic computing, it is proposed that the advent of Quantum Computing would give researchers the ability to scale optimally and hence lead to the creation of a truly intelligent machine.

 

Currently AI in its individualistic phases is everywhere. Most apps created now e.g. Airbnb, Facebook, Slack, LinkedIn all have the ability to learn a user’s preferences from the way he or she uses the app. It is said that the Facebook face recognition system trained on images on Facebook has a 98 percent recognition rate on over 1 billion people.

 

The main goals of AI research include reasoning, knowledge, planning, learning, natural language processing (communication), perception and the ability to move and manipulate objects. The ultimate goal of AI is the putting together of all these in one machine (Composite) and in essence achieve general intelligence like humans. This end goal is referred to as Strong AI

AI research is highly technical and specialized. Currently popular approaches include statistical methods, computational intelligence and traditional symbolic AI. There are a large number of tools used in AI, including versions of search and mathematical optimization, logic, methods based on probability and economics, and many others. The AI field is interdisciplinary, in which a number of sciences and professions converge, including Computer Science, Mathematics, Psychology, Linguistics, Philosophy and Neuroscience, as well as other specialized fields such as Artificial Psychology.

 

It is said that over the years, computing has moved from Desktop first approach to a Mobile first approach. It is proposed that in the coming years, any app that is to be taken seriously should have some form of AI, hence moving to an AI first approach. Companies such as Google, Microsoft, Facebook etc. have already made this transition.

 

 

H. A

Related Articles

Leave a Reply

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