blog imag

By Rishabh Arora

Machine Learning made easy for Human Learning

We are back with our 2nd article by Rishabh Arora in the Grandmother Speak series. Tune in as our in-house experts demystify Artificial Intelligence and Machine Learning concepts to simple, practical and snackable bites, just for you!

Humans learn using data consumed in the form of text, video, audio, etc. For example, when a child looks at a four-legged animal, it forms an image of it, records its name (e.g. dog), and stores it in memory. The next time when it sees a dog it knows what it is. If it says dog looking at a cow someone corrects the answer and says, it is a cow. The child stores the word cow against the new image. This forms new intelligence. This is how humans learn and train.

Machines also learn in the same manner, we feed machines data, tell them how and what to learn from that data to create intelligence. The machines use intelligence to give us answers. Based on whether the answer is right or wrong the machine learns more and forms new intelligence. Machines can also be made to learn unattended i.e. the process to learn and unlearn from data goes on without human intervention. All this is called Machine Learning.

Grandmother Example: When you say Alexa, play the latest news. Based on training, Alexa knows exactly what this request means. If, however, you say a word that Alexa has not heard before it tries to link it with the nearest possible match and gives an answer. If you cut it in between and ask the initial request again it knows its attempt to link was incorrect and it tries again. If, however, you go with the linkage and listen to what Alexa played it knows the linkage was correct and that becomes its new learning. This is Machine Learning.