PiMA
I’m skeptical of AI, but I still chose to attend a summer camp about it. My perspective is simple: disagree with knowledge, not from a distance.
Our team studied autoregressive models and WaveNet: how earlier samples inform the probability distribution of the next sample, allowing audio to be generated sequentially.
The team generalized the receptive-field relationship for different kernel sizes and geometric dilation rates, then proved it by induction. Mathematics reveals unexpected relationships, and the solver’s task is to find the root of that surprise. That was also why competitive mathematics gradually lost some of its spark for me: too often, everything already felt laid out.
How far back can one point “see”?
THE MATHEMATICS I KEPT
BEYOND THE MODEL
Though I could join the Southern Camp, I chose the Hanoi PiMA Summer Camp in the North.”
“After the first karaoke session, everyone seemed to understand one another more naturally.
Notes from people I met at PiMA
← BACK TO MATHEMATICS
CONTINUE TO ECONOMICS →