“What is Artificial Intelligence?” is one of the most asked questions of our time. With computers, software, DVR’s, cell phones, video games and so on, the use of technology in everything humanly possible has increased dramatically. Will this increase in technology eventually cause the death of the human race? Well there are some who believe so, and others who are afraid of the power that computers may soon possess.
Artificial intelligence is the ability of a machine to do tasks that would otherwise be beyond its capabilities. Unlike the natural human intelligence displayed on human beings and other animals, artificial intelligence is achieved by machines, not humans. The difference between the former and the latter category is often displayed in the acronym commonly used. IQ, Artificial intelligence, Whetesha, R&D, Corporate Research, Computer science etc.
AI stands for artificial intelligence technologies, meaning ways of achieving goal setting, decision making, task automation or general purpose machines. Such as Watson, the computer brain, and its array of programs that are already in use. These are just a small example of what is AI technology. And what we see today is a very complex system, made up of many different sub-programs that together perform a variety of tasks. The goal of such systems is to perform human tasks more efficiently and quickly.
Computers are now able to beat the grand champion chess player, Garry Kasparov in one game. Is it possible to program a machine learning system to think like a human? Yes, and there is a company in New York called ‘Deep Learning’. They have already created an artificial intelligence called the Natural Language Processing (NLP) system that uses language processing to achieve high level of accuracy in speech recognition. Deep Learning has already conquered one of the most difficult tasks in machine learning, speech recognition, beating the grand champion chess player Kasparov.
In the future we will have AI supercomputers that can take over basic processing unit, like a computer, and allow the computer to think, reason and do anything that a human could. Deep Learning uses deep networks to make a program, with the same software and hardware capacity as computers; it learns through reinforcement. It builds on the previous learning by applying it to new data. It learns from its past mistakes and by continually updating itself through new input. It also uses large scale distributed computing using the Hadoop Distributed Computing initiative.
Deep learning is not limited to speech recognition. It’s not even limited to natural language processing either. It’s goal is to combine these two areas, with speech recognition and natural language processing. This is called reinforcement learning. The best applications of artificial intelligence right now are self-driving cars, robotic aid, e-commerce websites, and internet banking.
Some think that computers will be able to beat the best human athletes in games such as football. But this is still very early days yet. We will still have to do more work, for example improving the game analysis tools, such as stats graphing tools, that are used by gamers to analyze their games. Similarly, artificially intelligent machines will not be able to beat the top poker players at poker. But by plugging these poker skills into a Deep Learning system, it will be possible to create a competent online casino player.
In other areas we see the same trend of capturing more of the utility value of natural language processing. The developers of Google, and Apple even had a chat recently about how the two companies are working together to bring natural language processing to the iPhone. So perhaps the term artificial intelligence has already lost its meaning. The progress is rapid and shows no sign of stopping. Artificial intelligence is here to stay and may well spawn more advanced algorithms to take on more complex jobs.