What Life Has Taught Me So Far · 03
On the difference between a graph and a person
I spend a lot of time looking at numbers. In AI research, there are datasets, percentages, accuracy scores, training samples, testing samples, and graphs. My poverty-prediction research, for example, involved training a CNN using thousands of satellite images and measuring how accurately the model could predict poverty levels.
On a computer screen, poverty can become a number.
Then I volunteered at a homeless shelter.
There, poverty wasn’t a number anymore. I was helping distribute food, water, blankets, hygiene supplies, and other basic things to people who actually needed them.
That difference stuck with me.
A computer can tell me that a problem exists. Data can show me how large the problem is. AI might even help us figure out where the problem is happening. But none of that completely tells me what it feels like to be the person living through it.
I think that is something I want to remember as I get deeper into AI. It is really easy to get excited about a 90% accuracy score and forget what the numbers might actually represent.
What life has taught me so far: Data can help me understand the world, but I never want to forget that there are actual people behind the numbers.