Natalie Feng

Why do you choose data science as major or career? This question is brought out frequently. To me, because machine learning mechanism applies to a broader context than a specific technical area.

Most subjects share the same rationale, I don’t think there exist crisp cuts between areas. Data science is a handy way to crack codes of such “universal rules” behind.

Machine learns, we human do the same. It is us who taught machines how to learn. For instance, when computers are trained to tell apples from oranges, one way is to differentiate apples that mostly look like oranges (orange-ly apple) with oranges that look like apples (apple-ly orange), another way is to distinguish typical apples that have nothing like oranges (apple-ly apple) with typical oranges (orange-ly orange). Both ways sound familiar with how ourselves learn, right?

It is because our perceptions of objective world is just like models that machine builds to do analysis and prediction. Although all perceptions and points of view are subjective, they can be trained to approximate reality as close as possible. To build the most close-to-objective models in our mind, two elements are required: speculation and experience. Speculation is model structure for machine, experience is equal to data points.

Reading, traveling, social interactions are major ways for us to collect data. Whereas thinking on top of if allows us to put puzzle pieces together. Digestion is the key. Experience is of no use without being processed. Vision is not only about how much one reads, travels, or being schooled, it is about what one extracts from those reading, traveling, schooling experience, etc.

To experience without thinking is blindness, to think without experiencing is idleness. I know people who traveled and experienced a lot but underwhelm the audience, this is where digestion is needed. One interesting example, studying foreign languages is beyond learning other ways to express, it kicks in to help us calibrate our understanding of all languages including the native tone. We are so used to certain words from our native languages that they have lost the authentic function to us, we automatically put an equal sign between the word and its designated meaning, without realizing it is only one projection of the subject which is objectively “unknown”. By digging into other expressions from new languages, more data points are collected, this allows us to calibrate our approximation to objectively express this unknow subject.

So education is about learning how to learn, beyond what we have learned, just like our mind needs being sharpened constantly, using thinking as the whetstone.