Humans have always done a super job of simplifying their lives with inventions dating back to the hunting spear. Hop skip and jump thousands of years, from cave to condo dwelling, and we are still looking for ways to get our job done – with speed and success.
Even in our very lifetime, we have witnessed the transformation of entire lifestyles – from the way we shop to the quality, and quantity, of our relationships.
Computer, and its computing abilities, have had a profound impact and we continue to find ways to make them more powerful, smarter, and more and more…like us.
Artificial Intelligence is that all-encompassing endeavour of enabling computers to think and act like humans. Why? So that computers are capable of learning, thinking, reasoning, problem solving, and making decisions like humans can, more consistently and with incredible success.
Intelligent computers who can mimic intelligent humans, if you will.
Despite years of research and attempts, there is no system that has managed to reach such a capability till date.
With a fully functional AI system still very much out of reach, a smaller and more successful aspect of AI has definitely made rounds to the winner circle – Machine Learning.
Machine Learning is a form of data mining under the broad field of data science. The latter is essentially the art of extracting information from large data sets.
In ML, the machine is capable of also learning from the data. So, in layman’s terms, it is the capability of a machine to take existing data and find meaning in the data to make useful and informed decisions.
People who deal with large data and creating intelligent algorithms to extract information from them are called data scientists. Data Science gives us the mathematical, statistical, and computational tools to function in AI space.
Drilling further down to more specialized super powers of computing systems, is Deep Learning. In DL, the computer learns to do a lot of human-like things – understand sound, text and images. It works somewhat like the human brain using up a complex system of neural networks.
Result? You get driverless cars, for instance, that are not just machines that have learnt how to drive without human assistance, but are also able to correct themselves when faced with a pedestrian, a traffic light or the sound of an ambulance, without hitting either one.
|Artificial Intelligence||Machine Learning||Deep Learning|
|Includes all systems that mimic human intelligence, including ML and DL based systems||Subset of AI||Subset of ML|
|Covers all aspects of cognitive behaviour in artificial systems.
– Narrow AI deals with one particular task done same or better than humans
– General AI deals with all tasks done the same as humans
– Strong AI deals with all tasks done better than humans
|ML deals with algorithms that get trained from past data to model a network to process new input.||DL is more sophisticated subset of ML that uses multiple layers of networks that link the output to input.|
|Uses smaller datasets and can be done on low-end machines||Uses large datasets and requires sophisticated machines|
|Ideally, a strong AI system should be advanced enough for no human intervention||Human intervention needed when the model makes inaccurate assessments||No human intervention needed as the complex system of neural networks is advanced enough to learn and improve on its own. Self- driven cars, for instance.|
For the layman learner, with real intelligence, the difference between ML and DL can be quickly illustrated by designing a hypothetical program that equips your toaster to grill you a sandwich whenever you are hungry.
You say “sandwich” and the toaster, that is able to recognize your voice, greets you with a “your wish is my command” and out comes a grilled cheesy goodness.
That is a highly intelligent toaster fed with some really good ML algos. It is also able to interpret words like food, grub, repast…
Now suppose you are having a really bad day. Your boss has singled you out to slave away on a project without a lunch break, you have a sore throat, and to top everything else you just stubbed your little toe on the kitchen table. You let out a squeal and say “I am too tired to cook!”
If your toaster is able to recognize your distress and grill you a pesto chicken sandwich to make you feel better, even without the command word, what you have is a super super toaster on DL fuel!
And both the super and super super toasters are AI.
In real life, AI applications have infiltrated a significant aspect of our lives, in businesses, sports, even politics. According to a study conducted by Research and Markets, in 2017, the market for ML is projected to grow by over 44% by 2022.
In investment speak, that is a jump from about 1.4 billion USD to nearly 9 billion. Highly skilled ML scientists are projected to make over $120,000 in salary, to begin with.
Companies, across the globe are opting for ML in their business practices to compensate for the lack of appropriately skilled employees. Plus, the advantages of consistency in performance also helps efficiency in the long run.
Manufacturing and allied corporations are spilling over with automation supported by AI and ML. Corporations, from small to large sized, are demanding resources related to cloud computing, storage, IoT devices, and all such enhancements that help in growth.
Examples of AI enabled upgrades in companies include Alexa for work at Amazon, where employees can use Alexa as their assistant to facilitate business proceedings. Google is employing natural language processing, voice search, image identification and other applications of ML. Netflix uses AI to optimize user experience by learning viewer habits through their viewing behaviour.
In politics, we have already witnessed and will see even more AI influence in targeting news to people. Call it news or fake news, the fact that modern and “smart politics” can use algorithms to introduce bias among populations and thus impact the process of democracy, goes to show the reach of AI.
In sports too, AI is being used as a tool to influence consumer engagement, through
– automated sports journalism, whereby hard data from sports can be converted into content,
– wearable AI Tech that allows to improve sports performance,
– video analysis to help detect fouls and thus auto-refereeing
Basically, AI can increase chances of fair play, enhance spectator experience, generate and analyse sports data for training, and introduce measures to help boost the sports industry as a whole.
The above examples are just a few instances of AI and its reach. With the world heading towards smart this and smart that, the intelligent choice would be to embrace the artificial one and understand how significantly our lives have become entangled with its applications.
If you are interested in a career in AI, explore your interests and follow our articles below, and beyond, to see how far you can go.
Good luck Mr. Anderson!