Do you know Machine Learning?

 Machine Learning is like Jesus, It's everywhere...

From pizzerias to Notco (a company which uses AI to make vegan food that tastes like meat) and from banks to Netflix all are using Machine Learning. But can machines actually learn something? 🧐

There are several algorithms that improve performance on a particular task with experience, that's it. By the way, if anyone asked, that was the definition of Machine Learning. The thing that computer systems can actually increase their performance or learn tasks is what AI is driven by. 

Machine Learning is basically divided into 3 categories, viz, Supervised Learning, Unsupervised Learning, and Reinforcement Learning. 

    Supervised Learning is learning from a training set of labeled examples provided by a knowledgeable external supervisor. Each example is a description of a situation together with a specification—the label—of the correct action the system should take to that situation, which is often to identify a category to which the situation belongs. The motive of this kind of learning is for the system to extrapolate, or generalize, its responses so that it acts correctly in situations not present in the training set.

    Unsupervised Learning is typically about finding structure hidden in collections of unlabeled data. And there is no supervisor. It is mostly used when you don't know what you are looking for but just wanna know what's hidden in the data.

    Reinforcement Learning is learning what to do when given a situation so that it can maximize the reward. Humans and even animal's way of learning closely resembles this kind. Whether we are learning to drive a car or to hold a conversation, we are acutely aware of how our environment responds to what we do, and we seek to influence what happens through our behavior. 

There is also another kind of ML called Semi-supervised learning. Some Algorithms deal with data that is partially labeled. 

Machine Learning consists of algorithms whose basic element resembles neurons in a brain. The subset of ML which deals with neural networks is called Deep Learning. 

Some may find this unusually straightforward, they might think.. Oh! I just have to write this algorithm and it will learn on its own... Wait till you find out about 'hyperparameters'. 

If it was that easy, a robot would have been reading this blog. 😏

Machine Learning requires a lot of computing power. You'll know when you start building a model. That was the reason it was underrated all these decades. In this age, computing power might seem enough but the way the research goes even the best GPUs (Yes... GPUs are used for training ML models coz they can process Matrices or Images faster than a CPU) will become obsolete in the coming years.  

ML is also categorized based on whether they can learn incrementally or on the fly, viz. Batch Learning (Data is fed in batches and a fully trained model is deployed and stops learning) and Online Learning (System is trained incrementally by feeding it data individually or in mini-batches, system can learn about new data on the fly).

One more way to categorize ML is by how they generalize. There are 2 approaches, instance-based learning, and model-based learning. 

Instance-based learning: the most trivial form of learning is simple to learn by heart. If you were to create a spam filter this way, it would just flag all emails that are identical to emails that have already been flagged by users.

Model-based learning: to generalize from a set of examples is to build a model of these examples and then use that model to make predictions.

PS. Remember the time when an annoying kid who just doesn't get it and the teacher looks pissed... Well in this case you are the teacher and the student is a brainless box.