All that is under the Sun
Imagine you’ve built an AI model to sort fruit. After training, you test it on 100 pieces of fruit (80 apples and 20 oranges). The model correctly identifies 85 of them. Is it a good model? Your first instinct might be to say it’s “85% accurate,” and therefore pretty good. But what if I told …
The Art of Choosing One From Many: Categorical Cross-Entropy and the AI’s Grand Decision Imagine you’re a librarian training a new assistant. This isn’t just any library; it has 50,000 different sections. You hand the assistant a book and ask, “Where does this go?” The assistant, being new, doesn’t just point to one section. Instead, …
Imagine you’re trying to teach a machine the difference between a cat and a dog. You show it a picture of a cat, and it says, with 90% confidence, “That’s a cat.” Great! Now you show it a dog, and it says, with 60% confidence, “That’s a cat.” Not so great. The second guess is …
Imagine you’re teaching an archery student. They loose an arrow, and it thuds into the target, a little off-center. You wouldn’t just say, “You missed.” To be a good teacher, you need to know how much they missed by. Was it a near-perfect shot, just a hair’s breadth from the bullseye? Or did it land …
“Reality is that which, when you stop believing in it, doesn’t go away.” Philip K. Dick In the world of Large Language Models, the line between generated reality and objective reality can often feel perilously thin. We, the builders and researchers, stand at a fascinating and frustrating frontier. Our creations can draft elegant poetry, write …
Welcome back to All that is under the Sun! In our last blog post in Learning AI Series, we unmasked “Optimization,” the tireless engine that helps AI models strive for their “best selves.” We saw how AI isn’t just born brilliant; it learns and improves through a process of iterative refinement. But that begs a …