Like many people, I’ve learned more about Artificial Intelligence (AI), from ChatGPT to Llama. Large Language Models (LLM’s) have become much better at a certain kind of pattern matching.
It made me think about a project I participated in several years ago. Galaxy Zoo began as a way for thousands of volunteers to help astronomers classify galaxies.
I was inspired by the idea that I could contribute to cutting edge scientific research just by looking at pictures of galaxies and indicating which ones looked like spiral or elliptical galaxies. Almost anyone could do this!
There was too much data for astronomers to look at and classify every single galaxy. Because humans are good at pattern matching, it made sense to get many volunteers to help out with the effort. Where the volunteers disagreed, an astronomer could take a closer look.
Today, I found myself asking the following question.
Is it possible for AI to do what thousands of volunteers were able to do a decade ago? And now, I have an affirmative answer.
For Galaxy Zoo’s Euclid project, volunteers today can help train AI algorithms to classify even more galaxies!
It’s fascinating to see the convergence of crowd sourcing and AI.
Leave a Reply