Machine Learning, we all have heard it and heard it ears full. Yet we hesitate to get a hang of it. “The only stupid question is the one you don’t ask” So let’s ask a few fundamental questions Is Machine Rocket Science? Q) Learning No, it is used in Rocket Science though. Ans) Why is it that we are scared to take a peek into it? Q) Maybe what it does seem like a miracle to us. So we assume it is something out of our scope of learning/understanding. Ans) How tough/complex is it? Q) Anyone who has dared to fight this hydra knows that it is a child’s play (well that was an understatement but you get the idea). Ans) So what is it? Q) It is an attempt to make things more intelligent. Most of us have come across terms like “Artificial Neural ”, it is an attempt to replicate the working of the human brain. Even something like this is not necessarily always complex. At its heart, it is just multiplication and differentiation. Yes, Maths at it again but it’s rather what you learned at school, no different (This coming from a guy who is petrified of maths). Ans) Networks What does intelligent mean? Q) Understanding concepts or patterns behind the working of something. It could be understanding Emotions, making sense out of Human Languages (Ex: English , Hindi , French) and cool stuff like predictions. Ans) So what can it do? Q) Well Everything that a human can and a lot more. Some applications are really (Really REALLY !!) cool. Ans) Ok …. ? Like what? Q) Consider Following Ans) Like out of a billion choices on eCommerce websites. 1) predicting most relevant option Remember ? Well for all those who have found a Hot Match, thank you Machine Learning! 2) Tinder uses it to guess your mood and recommend the movie that you will be most interested in. 3) Netflix uses it to guess the out of a billion (even a few hundred billion) results. 4) Google most relevant page It is being used in to like Cancer before a person actually gets infected by it. Goosebumps anyone? 5) Medical field predict diseases My personal favorite : and type 6) Cortana Siri language understanding bots. 7) Everything!! :D Now let’s dive a bit deeper Netflix It has a genre tagged to a respective movie. Example : Star Wars is tagged adventure (OFC it is Aventure!). It also has a few other tags like actors, director, production houses, description, runtime etc. Now when you watch a movie, it records all the above info plus some extra info depending upon your reaction. Reactions like : Now it finds patterns in your behaviour. How much of it did you watch? How many times did you pause it? So it funnels down the result something like this: You like X genre -> 100 options You like Y actors -> 50 options You don’t like very long movies -> 10 options You mostly prefer animated movies -> 5 options Now these 5 options are the recommendations it will pitch at you but it doesn’t stop just there. You usually watch movies between 6 P.M to 10 P.M -> Schedules recommendation You usually watch scary movies before sleeping -> Prioritizes scary movies near 10 P.M slot Google (Text Analysis) Everything from the suggestions that Google displays when you start writing a text to the actual results that Google displays use machine learning. It uses NLP or Natural Language Processing. Natural languages are the languages humans use to communicate with each other. It understands language by converting text into vectors. (Yes the concept perplexed me too the first time I heard it) . N depends usually and roughly on the number of rules in a language under analysis. Example English is inferred to have rules between 300–400. So every variable in the matrix points towards a rule. Think of word vector as a matrix of size N Now the question is what value should be given to which rule? Q) Ans) I don’t know! :D Why am I so excited about not knowing the answer? Q) Because this is the power of Machine learning! Ans) It automates this process. These . means the vectors contain semantic meaning Semantic context . Results almost made me do the Archimedes. : Consider 3 sentences Example Messi scored a goal Ronaldo missed the last penalty Mukul missed his sleep Now the traditional learning would infer that sentence 2 and 3 have a same word ‘missed’. Rest no similarities. So 2 and 3 are closer. Stupid, right? Whereas our brain knows that 1 and 2 are used in same context which is sports or football precisely. BTW so do our Vectors ;) The vector of Ronaldo will have a value much closer to the vector of Messi. So when we find the similarities between the sentences using vectors we get 1 and 2 are the closer ones. Smarter, right? So a matrix of numbers can understand the language and the context? :O Q) , Ans) Yes rainbows in your eyes and wide open mouth are normal at this point ;)
