You can’t conclude a discussion on Machine Learning without mentioning classification. Classification is a machine learning technique where the machine is trained to predict the label of the given input data.
Alright, let’s cut the jargon and get some real-world examples. Oranges and Bananas.
Let’s assume that we have a box of fruits that contain some oranges and some bananas. You are asked to pick one fruit at random and tell if it is an orange or a banana.
Pretty basic, right?
For us, it is straightforward. We would know the answer at first sight. But, how would a computer be able to tell the difference?
In classification, the machine would first be trained on some pre-labeled data. It would be shown an orange and we would tell it that the fruit is an orange. The machine would study the orange and remember its features - orange color and round shape. Then it would be shown a banana and the process is repeated. What are these features? A feature is anything that helps us uniquely label the data. In our example, the color, size, shape - all could be features.
Oranges are round in shape and orange in color. Bananas are elongated and yellow in color. When the computer is given a random fruit, it would check the features of the fruit. If the fruit is round in shape and has an orangish hue, it might label the fruit as an orange.
Now, since we have a basic idea of what classification is, let us find out how we apply similar concepts in our lives. Our lives are full of labels. We get labeled based on our faith, our place of birth, the language we speak, our political affiliations, or even our generic belief systems. Though such labeling has always been there, the rise in usage of the same has shot up with the advent of social media.
Social media is where people from diverse backgrounds converge. It is only natural that a difference in opinion would ensue. However, labeling (and trying to degrade) people who have engaged in a debate with you on social media platforms based on the afore-mentioned features has been a disturbing trend of late.
Suppose that news channel ABC posted an update about a key announcement by the government. Under that post, your friend X wrote a comment questioning the logic behind the latest announcement. Unsurprisingly, many replies would spring up under this comment, most interested in shutting your friend down and assigning them a label rather than actually discussing the point raised. Now, the label to be assigned to your friend depends on some features, the most prominent one being their name.
If X has a name that resembles that of the majority community, they are most likely to be labeled as a “sickular liberal” (Yes! New words are being invented). On the other hand, if X is from a minority community, chances are that they would be straight up labeled an “anti-national”.
Limiting labeling to the social media universe seems unfair since the practice is religiously followed outside of the virtual world as well. ‘Outspoken’ and ‘arrogant’ are common labels used to describe children who have a different opinion to the elders of the house.
A popular female actor, who would traditionally be labeled ‘item’ and ‘bomb’ by the misogynistic mob, would easily be labeled a ‘feminazi’ the moment she decides to display her mind on the screen instead of skin.
What exactly is wrong with labeling? Labeling, inherently, is not wrong. But when we use labels to counter a point raised by a person, we reduce them to just that word. The questions raised by the person are then seen as politically motivated ramblings of the group represented by the label. This hinders healthy discussions and promotes bias.
How cool would it have been if we all looked at the points asked objectively without doting on the person who asked the question? Would that be a reality ever? I don’t know. I am agnostic!
Machine Unlearning is a series broken up into tiny, one-minute readable pieces to humor our ever-shortening attention span. Sharing the links to every single piece right below: