Decoding Artificial Intelligence (AI) & Machine Learning (ML)
In the current scenario AI and ML have become buzzwords but they are often used interchangeably. It is important to break these words down to better comprehend their applications.
AI- It means getting a system (computer) to mimic behaviour of humans in a particular manner.
ML – Basically ML is a subset of Artificial Intelligence, contains the technique that enables computer to figure out from the data and deliver AI Application.
Above breakdown would have briefly given us a broader understanding of these concepts. Let us now dive deeper into their origins.
Artificial Intelligence was founded way back in 1956 as an academic discipline. The focus was on how to get computers to do the work considered as uniquely Human, which will required intelligence. As a starting point, many researchers focused on simple problems i.e. solving logical questions or playing checkers. Many times when they we are playing games (Rummy or Ludo) against a computer as an opponent, some kind of “artificial intelligence” experience is observed in the moves of the computer, especially when we lose to them.
Gartner Glossary explains Artificial Intelligence (AI) as “applying advanced analysis and logic-based techniques, including machine learning, to interpret events, support and automate decisions, and take action.” We can describe AI in today’s context as, collecting data, take the learning from it and redeploying outputs that helps us in finding solutions to the problems.
Machine Learning is based on Algorithms that can be learnt from data without relying on rules-based programming (Dorian Pyle and Cristina san Jose, Mckinsey consulting).
Let me explain this concept with a very simple example which I am pretty sure that we must have gone through in our childhood. When I wanted to learn on how to ride a bicycle, our dad didn’t have a rule book on mastering this joyful experience. I were simply put on a bicycle, was asked to start paddling along with a gentle push from behind by my dad. As I started riding, I was continuously collecting data on the way i.e. how balancing is working, when to put the break or speed up or down as well understanding and absorbing learnings from my every crash or fall.
Machine Learning is very similar to the above example, where we simply take the data, decide what we want to do and work with/on it and see the desired outcome.
Gartner Glossary defines Machine Learning (ML) as “Advanced machine learning algorithms are composed of many technologies (such as deep learning, neural networks, and natural language processing), used in unsupervised and supervised learning, that operate guided by lessons from existing information.” ML can be called as a subset of AI, and although all ML counts as AI, but not all AI counts as ML; instead, ML learns over time.
- AI & ML was around for a very long period of time, but the major challenge was lack of data collected to apply AI or further precisely to, ML. Data now has assumed significant importance and with the changing times organisations are understanding the true power of data collection and its applications.
- ML needs neural networks which requires high intensity computers to do the processing parallelly.
- AI & ML can be considered as magic words as they can’t solve all the problems of businesses, but they are certainly the most in demand tools which are shaping the way in which organisations take decisions.