Part 7/12:
Shapiro shares his experience participating in Kaggle competitions with evolutionary algorithms that evolved neural network structures, highlighting the value of experimenting and deepening understanding through creation. The goal is to internalize the workings of neural networks to an extent that allows creative experimentation and problem-specific customization.
Step 4: Engage in Competitions
Next, he recommends entering competitions like Kaggle. These challenges simulate messy, real-world problems and force learners to adapt to noisy data and unclear goals. Though he admits he never aimed to win competitions, participating was instrumental in accelerating growth.