Popular Mechanics (South Africa)
How to be good at what you do:
Azalia Mirhoseini, senior research scientist at Google Brain.
(1) What better source of inspiration than the human brain? Deep learning, which is only vaguely inspired by, but still far from, how the human brain functions, is already revolutionising the world. (2) We spend most of our time on the journey. We better be having fun doing so. (3) Of course, caffeinated! That’s the whole point! (4) Just like the majority of people in the field, I program my deep-learning algorithms in TensorFlow and use its Python API. Python is really the established language of AI. (5) I work in an open office space, so a pair of noise-cancelling headphones is a must. (6) Reinforcement learning is more creative. Something I find intriguing about RL is that it can solve decision-making problems in ways we humans can’t think of (just like it solved chess and Go). We’re going to see revolutionising applications of it in the coming years. (7) Innovation is what unlocks the future.