Original article was published on Artificial Intelligence on Medium
What is the difference between AI, ML, and RPA?
What is the difference between RPA, AI, and ML?
How much of AI is too much of AI?
How much of AI is AI? The media noise has distorted the idea of AI so much that catchphrases, pop media, books, and movies paint droids luxating in the metallic gleam of semi-viscous liquids or the cavalry of droids that are set to annex countries, accumulating at the city gates. The line of the divide is very vague, sometimes, like how essential it is to differentiate in a surgical device the measure of true human movement, and human augmentation.
Even though the term “artificial intelligence” has undeniably been misused, the technology is now deployed in healthcare, warfare, developing books and music, scrutinising and skimming through resumes, determining your creditworthiness, modifying photos on your phone, and websites with AI-powered chatbots.
The latest in the distortion line was the Oral — B’s Genius X toothbrush, with supposed AI abilities. Though the toothbrush uses sensors that provide feedback on whether you are brushing for the right amount of time, dubbing its artificial intelligence is trivial.
The sci –fi version of a droid many times smarter than a human, what experts call artificial general intelligence, will happen, if we are to create one, only a long way in the future.
Google search, image recognition software, personal assistants, like Siri and Google, self-driving cars, and IBM’s Watson are examples of AI that incorporate algorithms and provide an output. However, these systems do not have any understanding or awareness of what it does.
An awareness of AI and its related technologies, machine learning and deep learning, help you to understand customer service processes and the company’s time and money.
Can we teach machines or do machines learn?
Machine learning, a subset and application of AI, endows a system with the ability to learn, identify, and make predictions using data. Instead of using code to define a problem, a system can now be taught to learn whether some fruit is an orange or apple, or an animal is a cat or a dog.
Illustrating the learning curve of a machine
So, you are trying to create a program that works toward recognizing pets. You do this the old and traditional way by programming, mapping clear-cut rules, such as “pointy ears” or “perky eyes.” But the catch is when you show the picture of a leopard or a tiger, and then you define the co-ordinates for a cat, but how does it tally when you show the image of possums or animals with similar size to that of cats. For this, you would have to define all sorts of permutations and combinations ranging from furriness to sharpness to all sorts of twistiness. At this stage, you almost wish that it would have been better if the machine could learn itself. It is better to give the machine a huge collection of cat photos and let it decide for itself on the patterns it finds in what it sees.
Over time, the machine, will connect the dots, and it will get better with each and every version. The single most advantage of this method is that you do not have to program it. In spite of all the tinkering that you do to modify and process the system, you still don’t tell the system what to look for, but you give data 1s and 0s, letting the system figure and identify patterns that even humans are incapable of.
What is the difference between AI and ML, then?
The basic difference is that artificial intelligence uses algorithms that can work with its own intelligence, aping human intelligence, while machine learning is about extracting knowledge from data.
Artificial intelligence has a wider scope of application, and it can be used to implement and perform complex tasks while machine learning has a limitation in that it is only trained to perform a certain subset of tasks. The applications of artificial intelligence are in Siri and customer support using chatbots, while machine learning is found in online recommender systems, google search algorithms, and Facebook auto-tagging suggestions.
Then, where does robotic process automation stand?
With all this comes a sort of bafflement as to what is the difference between artificial intelligence (AI) and machine learning (ML) and robotic process automation services (RPA)? Simply put, robotic process automation is a software robot, following strict rules, to perform repetitive tasks. Therefore, RPA is more like a staffer at an organisation performing well at his repetitive clerical job.