Why do Xavier or He initializations matter?
PICTURE THIS: DATA SPLIT
Fit on train, tune on val, report on test once.
Simple meaning
Random weights that are too large explode activations
WHY — Neural Nets instead of guessing?
Why interviewers care about Neural Nets:
on Neural Nets.
the situation, the default choice, and one exception - that reads as experience.
Name the idea, why it exists, then one short example.
End with when you use it and one common pitfall.
STEPS — What happens step by step?
Before you speak the answer, walk the interviewer through these steps:
- 1Random weights that are
too large explode activations
- 2Why it exists
too small vanish them.
- 3Xavier matches variance for
tanh or sigmoid
- 4He scales ReLU by
accounting for the half of pre-activations that get zeroed.
- 5Bad init is still
a common reason a deep net refuses to train.
- 6Close
When you pick this over the alternative.
EXAMPLE — See it in action
Here's a short line you can speak, broken into clear beats:
Note: Adapt this scaffold to your own project — keep it under 60–90 seconds.
Key takeaway
Random weights that are too large explode activations too small vanish them.